Next Article in Journal
Variation Pattern and Genome-Wide Association Study of Leaf Phenotypic Traits among Ancient Ginkgo biloba L. Populations
Previous Article in Journal
Understanding Stakeholders’ Perspectives on the Indonesia–EU FLEGT-VPA Scheme Implementation
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Forests
Volume 13
Issue 11
10.3390/f13111763
forests-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Forest Road Wearing Course Damage Assessment Possibilities with Different Types of Laser Scanning Methods including New iPhone LiDAR Scanning Apps

by
Tomáš Mikita
1,
Dominika Krausková
1,
Petr Hrůza
2,*,
Miloš Cibulka
1 and
Zdeněk Patočka
1
1
Department of Forest Management and Applied Geoinformatics, Faculty of Forestry and Wood Technology, Mendel University in Brno, Zemědělská 3, 613 00 Brno, Czech Republic
2
Department of Landscape Management, Faculty of Forestry and Wood Technology, Mendel University in Brno, Zemědělská 3, 613 00 Brno, Czech Republic
*
Author to whom correspondence should be addressed.
Forests 2022, 13(11), 1763; https://doi.org/10.3390/f13111763
Submission received: 30 September 2022 / Revised: 21 October 2022 / Accepted: 24 October 2022 / Published: 26 October 2022
(This article belongs to the Section Forest Operations and Engineering)
Browse Figures
Versions Notes

Abstract

:
Forests make up 34.1% of the Czech Republic total area and forest roads account for nearly the same length (47,465 km) as all other roads administered by the state and its regions (55,738 km). Forest roads are not as intensively used as other roads. On the other hand, as logging trucks carry the maximum permitted load on roads and forests create a specific microclimate, forest roads are subject to rapid wear. A road wearing course is generally designed for 20 years of service and for a maximum damage level of 25% before they are supposed to be reconstructed. To ensure this life cycle is adhered to, more efficient, faster, and more flexible surface damage detection adaptable for forest environment is needed. As smartphones and their optical devices, i.e., new iPhones with LiDAR sensors, become more advanced, the option arises to perform laser scanning on road surfaces using smartphones applications. This work aimed to test this technology and its precision applicability to assessing damage to a forest wearing course and compare it with another hand-held personal laser scanner (PLShh), represented in this study by GeoSLAM ZEB Horizon scanner, and more precise terrestrial laser scanning (TLS) technology, represented in this study by Faro Focus 3D laser scanner, which have started to replace tacheometric wearing course damage surveying thanks to their greater precision. So, this study involved a comparison of three alternative laser scanning methods focused especially on these, which are implemented in new iPhones for tacheometric surveying. First, a Faro Focus 3D laser scanner was used for the TLS method. Second, the PLShh method was tested on a GeoSLAM ZEB Horizon scanner. Third, another PLShh method using an iPhone 13 Pro with applications 3D Scanner and Polycam was evaluated. If we are comparing positional height accuracy of PLShh to tacheometric surveying on reference cross position height coordinates, ZEB Horizon achieved devXY and devZ RMSE 0.108 m; 0.025 m; iPhone 13 Pro with 3D Scanner app devXY and devZ RMSE 0.185 m; 0.021 m, and with Polycam app devXY and devZ RMSE 0.31 m; 0.045. TLS achieved the best results with devXY RMSE 0.049 and devZ RMSE 0.0077. The results confirm that only the TLS scanner achieves precision values in height differences applicable for an assessment of forest road wearing course damage measurement comparable with tacheometric surveying. Surprisingly, comparing the PLShh scanners to the TLS technology, they achieved interesting results, comparing their transverse profiles and 3D objects as digital surface models (DSM) of the road to TLS in height position. In transverse profiles, ZEB Horizon achieved devZ RMSE 0.032 m; iPhone 13 Pro with 3D Scanner app devZ RMSE 0.017 m, and with Polycam app devZ RMSE 0.041 m compared to the TLS method measured using a Faro Focus 3D static laser scanner. Comparing forest road DSM to Faro Focus 3D, ZEB Horizon achieved devZ RMSE 0.028 m; iPhone 13 Pro with 3D Scanner app devZ RMSE 0.018 m and with Polycam devZ RMSE 0.041 m. These results in height differences show that the height accuracy of PLShh achieves precision, which is applicable to determining the current shape of forest road wearing course compared to the required roof shape gradient. However, further testing provided the insight that such a kind of PLShh measurement is still only possible to use for the identification of a transverse profile shape, as in length measurement the length error increases. All PLShh are able to capture the current shape of forest road cross profile, but still they cannot be used for any design or calculation of material measurement needed for wearing course repair.

1. Introduction

Roads are essential for efficient forest management [1]. From a regulatory point of view, in the Czech Republic Roads Act No. 13/1997 Coll. [2], forest roads are no different from other roads. The same traffic rules apply, and they have to take the same vehicles. However, according to the Forest Act No. 289/1995 Coll. [3] forest owners can ask the state administration to restrict public access on the forest roads in the interest of forest management. The main forest roads, which are designed to be used all year round, are constructed with compacted sealed wearing courses, so the construction technologies employed and their cost are comparable with other types of roads. However, due to the low traffic on forest roads and the high cost of design and construction, repairs are often neglected. For these reasons, maintenance is key to managing a forest road network [4]. One of the problems encountered in a forest road wearing course is rutting [5]. The water flowing in the ruts concentrates and causes erosion rills. In other cases, it remains on the road surface and, if frozen, causes damage and the gradual disintegration of the wearing course [6,7]. Forest roads should always have a stable and high-quality surface for safe and comfortable driving, for which regular maintenance and repairs are required [5]. Monitoring the road surface cross-section allows damage to be repaired early and thus avoid higher future costs [8]. At present, a geodetic terrestrial method, namely tacheometric surveying using a total station with a global navigation satellite system (GNSS), is most commonly used to determine the extent of damage to a forest road wearing course. This is an accurate but time-consuming method, and the work in the field in particular is very lengthy. This is the reason faster methods are being searched for, tested, and verified. These should gradually replace geodetic surveying methods and facilitate the mapping of the current state of a large number of damaged forest roads. The assessment of forest road wearing course damage requires accurate observations of the road surface topography. As laser surface scanning seems to be promising in the long term [9,10,11], this paper focuses on this technology. Although image processing methods also achieve very good results [12], laser scanning achieves more accurate results and, above all, does not require the targeting of a large number of ground reference points (GCPs).
Large-scale mapping from the air, i.e., airborne laser scanning (ALS) [13,14,15,16], has proven to be insufficiently detailed despite efforts to refine it [17], and replacing ALS with unmanned aerial vehicles (UAVs) or drones [18] has proven troublesome in the case of forest stands [19]. The use of airborne ALS for forest road surface mapping is limited by the lack of point cloud density. In the case of UAV are disadvantages given also by the legislative restrictions, since the use of UAV for this purpose in the Czech Republic requires a flight permit and pilot registration, which is issued by the civil aviation authority. This fact, together with the risk of a plane crash due to collision with the canopy over the forest road, greatly complicates the procedure [19]. Therefore, attention has turned to terrestrial static laser scanning (TLS), but the method of measuring in the field is close to tacheometric surveying in terms of processing in terrain and similarly time-consuming. Hence, this is the reason why is this technology starting to be transitioned from static to mobile or personal laser scanning process. Mobile laser scanning (MLS) proved to be quite inaccurate in the forest environment [19] and the acquisition of the entire MLS system placed on the vehicle can be quite costly for most forest owners. On the other hand, hand-held personal laser scanning (PLShh) is generally showing to be more practical and less time-consuming than TLS and more suitable for immediate and accurate forest condition observation than ALS. Their use is more flexible, meeting current needs to determine the condition of forest roads and less costly. Completely new possibilities for PLShh are offered by innovations of smart phones using LiDAR technologies [20].
The main advantage of terrestrial laser scanning is that its view is not obscured by tree crowns, as is the case with ALS. Another advantage of TLS is the fast and easy, contactless collection of highly accurate data with very high resolution obtained in the optimum quantity/time ratio (thousands of points per second) [21]. Thanks to these advantages, especially high-precision data collection, TLS can also be used during construction to monitor road surface wear [5]. However, studies [22,23] that dealt with the evaluation of road surface and shape irregularities by the TLS method were mostly carried out in locations without forest stands nearby. Despite finding that the data can serve multiple purposes, including validation of other design parameters, such as transverse and longitudinal slope, they concluded that TLS still required more processing and training [23].
Static TLS technology has not been very successful in practice, mainly because a terrestrial laser scanner and its data processing software are more expensive than a total station for geodetic surveying. Moreover, a static terrestrial scanner needs to be moved about in terrain just like a total station.
Mobile laser scanning systems (MLSs) are widely used, especially in urban areas. The process is quick and easy and allows very dense point clouds to be obtained which accurately represent reality. They are widely used in cities for traffic management and parking systems or when assessing the technical condition of road structures. For example, Jaakkola et al. [24] introduced automatic methods for evaluating road safety by analyzing road cracks [25] and detecting road roughness [26]. According to Yan et al. [27], an MLS provides information on technical elements on roads accurately and efficiently. However, their study on the detection and classification of pole-like road objects using an MLS was conducted on a highway, where shading by surrounding objects or vegetation was not to be expected. Vallet and Papelard [28] state that MLS achieves much higher accuracy and density than ALS. Their findings were in urban areas and for transport use. Kukko et al. [29] presented a multiplatform MLS solution for map applications that require mobility in different terrains and river environments, and reliably created point clouds with high density and accuracy. At the same time, Kukko et al. in their another study [30] published an article on the use of MLS for monitoring stand condition and forest inventory. Although they managed to achieve high data accuracy, they stated that the absolute position of the data was highly dependent on the visibility of the Global Navigation Satellite System (GNSS) in the forest canopy.
Recently, a more practical data acquisition method for detailed surveying of objects has started to be used, namely hand-held personal laser scanning, referred to as PLShh in the case of the manual process. As stated by Gollob et al. [31], the use of new and modern sensors in auditing forest inventory has become increasingly efficient. Still, the most of forest inventory data are collected manually, by field surveys. The reason for this is the time-consuming process of data acquisition with static TLS and the potential incompleteness of the data obtained. These disadvantages can be overcome by PLShh. Balenović et al. [32] state that the recent rapid progress in sensor miniaturization has resulted in the development of lightweight hand-held PLShh systems; according to them, the difficulty of operating vehicles in the forest makes the use of human operators more practical.
Laser scanning using devices that generate 3D point clouds appears to be at least one direction in which data acquisition in forests will develop [33,34], especially hand-held personal laser scanners (PLShh). As PLShh are in this study meant laser scanning technologies which are carried in person as opposed to those carried on a mobile vehicle or means of transport, they are referred to in this chapter as MLS, but we understand that in some studies all moving systems are referred to as MLS. This paper aims to test this technology in determining the condition of forest road surfaces, with a focus on the use of the iPhone 13 Pro 2020 smartphones device, which uses a new built-in LiDAR sensor. Combining this with commonly available technologies would allow wide use by forestry personnel and full use in common practice. The device satisfies increasing demand for high-quality photographs from the public and professional users. As smartphone manufacturers take this demand into account, it can be expected that special apps will be developed with this hardware, which will ultimately be usable in the work process. Several small objects with known dimensions were scanned to test the accuracy and precision of the LiDAR sensor. Smartphones can be used not only for detecting 3D shapes in enclosed spaces, small areas [34], and open urban environments [25,27], but also on models of landscape terrain and structures in the forest, which are affected by the surrounding vegetation in terms of surface sensing. For example, in study [20], independent models of a coastal cliff were acquired with the ‘3d Scanner App’ using the iPhone’s LiDAR sensor, via SfM MVS photogrammetry, and with the ‘EveryPoint’ app combined with the iPhone’s LiDAR and camera photos. The study tested mobile apps 3D Scanner (Laan Consulting Corp, 2021 Laan Labs, New York, NY, USA) and Polycam (Polycam Inc., 2021 Polycam, Los Angeles, CA, USA) on an iPhone 13 Pro for larger continuous surface areas affected by the surrounding forest environment.
The goal of this work was to determine the accuracy and appropriateness of different laser scanning technologies used for surface detection, with a focus on suability for the identification of the technical condition of forest roads, especially those referred in this study as PLShh. A comparison of different sensors in their capacities for characterizing forest road surface topography reveals differences in their capacities in assessing road wearing course damage. This work tested technology of iPhone 13 Pro and its apps 3D Scanner and Polycam together with the other PLShh GeoSLAM ZEB Horizon and more precise TLS technology, represented in this study by Faro Focus 3D laser scanner, which is taken as a reference surface represented the current shape of forest road wearing course.
Accuracy verification was conducted to ensure that engineers and decision-makers can use mobile LiDAR technology fully as mobile LiDAR systems have their own limitations and their performance varies depending on several factors, such as range, object reflectivity, incident angle of laser pulse to the reflective object, and the accuracy of the navigation system used [20,25].

2. Materials and Methods

2.1. Experimental Design

The comparison of different scanning methods was carried out on a forest road “Skatulova” located in the Training Forest Enterprise “Masaryk Forest” Křtiny, approximately 20 km northeast of Brno (Figure 1). Forest road “Skatulova” is constructed with the wearing course from penetration macadam. The original shape of the transverse profile was roof-shape with 3% grade to each side. Wearing course was disintegrated and transformed to various shapes by haul-track transport after 20 years of service, which was captured by TLS Faro Focus 3D. According to our interim results, TLS Faro Focus 3D achieved applicable accuracy and was interchangeable with tacheometric surveying. Hence, we used its scanned surface as the reference wearing course shape.
Three damaged sections of the forest road, each with a length of 150 m, were selected on the road (Figure 1) according to its damage for the purpose of testing the accuracy of chosen laser scanning methods. The road sections were selected considering the different vertical and directional routing. Section 1 includes a sharp curve and grade, Section 2 includes a curve with a smaller radius and only a slight grade, and Section 3 is mostly straight and flat. On the sections of the road, 450 m in total length, reference crosses were marked (spray-painted) at each side of the road on its shoulders at 10 m intervals There were 32 reference crosses in each section (altogether 96 reference crosses). These crosses subsequently served as accuracy benchmarks (as reference values for the laser scanning methods accuracy and are in the article referred as reference crosses). Spatial coordinates (X, Y, Z) in centers of reference crosses on each section of the forest road were tacheometrically measured in the local coordinate system using a Trimble M3 total station, always from a single position in the middle of the section to avoid deviations caused by instrument relocation. The geodetic survey of the reference crosses of each section took about 30 min (all sections about 90 min). The time required for data collection, in the case of the TLS with the Faro Focus 3D scanner, was approximately 3 h (all sections 450 m long), in the case of GeoSLAM ZEB Horizon scanner it was about 20 min, and with iPhone Apps it was about 30 min together (Table 1).

2.2. Point Cloud Data Collection

The road sections were laser-scanned with the entire width of the road surface using different laser scanning instruments (static laser scanner Faro Focus 3D (Figure 2a), hand-held personal laser scanner ZEB Horizon produced by the GeoSLAM company (Figure 2b) and hand-held personal device smartphone iPhone 13 Pro with built-in LiDAR scanner (Figure 2c) using the 3D Scanner and Polycam apps).

2.2.1. Terrestrial Laser Scanning (TLS)

A Faro Focus 3D Scanner was used for the purpose of this measurement. This is a static panoramic scanner with phase-based distance measurement. The range of the scanner is from 0.6 m to 120 m, allowing it to record up to 976,000 points per second, with an accuracy of approximately 2 mm at 10 m. The scanner works with different resolution values, ranging from very low resolution (1/32) to full resolution (1/1). For this study a resolution of 1/2 was used to ensure a high density of points, creating approximately 3 mm spacing within 10 m of the scanner.
The scanner was gradually placed between selected and marked reference crosses so that the crosses could subsequently be identified in the resulting dense point cloud. In each section, 16 scans in total were performed in this way. Reference spheres serving to connect all scans to one point (set) cloud were placed at each scanning station. The diameter of the spheres was 20 cm. The scanning of one segment took about one hour (all sections 3 h), with resulting point cloud with 158,999,237 number of points, point density 139,473.01 per square meter.
The measured data were exported from the scanner into Faro Scene software. As a first step, faulty points were eliminated using a Stray filter, based on automatic filtration. The next step was to manually filter individual scans and combine them in one point (set) cloud based on the automatic recognition of reference spheres. Based on the report from the Faro Scene software, final point clouds of sections were aligned with a mean accuracy of about 4 mm. The data were recorded using the local coordinate system, so there was no need for further transformation to coordinate systems. The resulting point clouds were exported to. xyz format.

2.2.2. Hand-Held Personal Laser Scanning (PLShh) with GeoSLAM ZEB Horizon

Data were acquired using a GeoSLAM ZEB Horizon scanner, which was launched by GeoSLAM Ltd. in 2018. Thanks to its simple and compact design, the data acquisition process was easy and fast. With a range of 100 m, it is suitable for both indoor and outdoor measurements. It collects 300,000 points per second with an accuracy of 1–3 cm. The scanner uses a SLAM algorithm (simultaneous localization and mapping), which allows it to gather data from sensors, create an image of the surrounding environment, and locate their position within it. When processing data, the algorithm gradually connects individual scans with the data from an inertial unit with a trajectory. This file contains location information for each scan at any given time. Thanks to this technology, scanning was carried out continuously, without the need for placing the reference spheres.
Each section of the forest road was scanned individually, with a walking speed of approximately 4 km/h. First, the sections were scanned along the right side and then on the way back along the left side, so that the scanning began and ended at the same spot. This method guaranteed higher point density and better visibility of spray-painted reference crosses in the resulting point cloud. The scanning of one segment took about six minutes (all sections approximately 20 min together), with the resulting point cloud of 29,838,103 points and point density of 14,048.07 per square meter.
The data obtained from the scanner were processed in GeoSLAM Connect software. As a result of importing the dataset to this software, trajectory files and .las files representing point clouds of all three sections of the road were processed. The resulting .las files were then exported to CloudCompare software where a further process were carried out. The first step consisted of classifying the ground points and non-ground points of each section using the CANUPO plugin which automatically separates the point cloud into two subsets. The next step was to clean the ground point cloud using the SOR filter and Noise filter tools. These tools were used to clean the ground point cloud, removing points by a selected mean distance and radius of filtering noise and removing isolated points.

2.2.3. Hand-Held Personal Laser Scanning (PLShh) with iPhone 13 Pro

Another method was adopted using an iPhone 13 Pro. The LiDAR Scanner measures the distance to surrounding objects up to 5 m away, works both indoors and outdoors, and operates at the photon level at nano-second speeds [25]. The forest road sections were scanned using iPhone apps 3D Scanner (Laan Labs, New York, NY, USA) and Polycam (Polycam Inc., 2021 Polycam, Los Angeles, CA, USA), both of which allow mesh and point clouds to be colored and exported. Scanning with the iPhone 13 Pro was done as with the GeoSLAM ZEB Horizon, but only in one direction. It was necessary to proceed at a speed of 3 km/h in the middle of the road and continuously monitor the surface model creation on the smartphone display.
The resulting data were then exported from the device in an .las file to CloudCompare software. No further processing or filtration was needed because the exported data represented the ground points of the scanned sections. The scanning of one section took about 10 min (3 section 30 min together) with the resulting point cloud for 3DScanner App having 19,577,813 points and point density 17,294.89 per square meter, and for Polycam App with 16,373,280 points and point density 14,993.85 per square meter.

2.3. Point Cloud Data Processing

Data from various devices exported in .las format were processed in open-source software CloudCompare. First, in each point cloud from all sections, centers of reference crosses were identified based on color (in the case of Faro Scanner and iPhone apps) or intensity (GeoSLAM Horizon) in the form of points and saved in .shp format.
This means 96 points were identified in all sections and in each point cloud from different laser scanning methods. These points were manually transformed using the Align tool to points taken using the total station. The scale adjustment option was not used during the transformation, so the points and subsequently the point clouds retained their original size and shape. The coordinates of the aligned points were exported in .txt format and the positional location X, Y and height Z of the transformation was evaluated in MS Excel software. The transformation matrices for each device were used and each section were saved separately and then applied to the appropriate point clouds. This served for comparison of the positional and height accuracy of each scanning method with the reference values of measurements using a total station. Further accuracy evaluation of each laser scanning methods was performed using ArcGIS Pro software.
Afterwards, from each laser scanning, data were processing in the 3D objects as digital surface models (DSM) of the road (Figure 3). Point clouds for all sections and all methods were interpolated into raster DSM with resolution of 0.01 m by LAS Dataset To Raster tool in ArcGIS Pro software (Figure 3). Subsequently, the lengths between the crosses in the direction of the path on the left and right side of the path were measured on all generated DSMs (for data from total station and for all methods). To determinate the transverse accuracy, transverse profiles were vectorized from the reference crosses for all methods and all sections (together 48 transverse profiles) (Figure 4).

2.4. Accuracy Asessment

The accuracy of the different scanning methods was compared in several ways:
1. The accuracy of the static laser scanner Faro Focus 3D, hand-held personal laser scanner ZEB Horizon and hand-held personal device iPhone 13 Pro using the 3D Scanner and Polycam apps was compared with the position and height of reference crosses measured by the tacheometric surveying. In each section, there were 32 reference crosses, for a total of 96 reference crosses.
2. The lengths of individual sections, again based on identified reference crosses on both sides of the road were compared with reference crosses surveyed by tacheometric surveying. The lengths of each section were measured on both sides of the road and the aim was to determine the change in length with increasing distance from the beginning of the section. The results were calculated from six longitudinal profiles, created on both sides of the road (Figure 5).
3. The lengths between the identified reference crosses within transverse profile were compared with the lengths measured between the points measured by tacheometric surveying. In each section, 16 transverse profiles were measured, in total 48 transverse profiles.
4. Using the Stack Profile tool, the heights from created DSMs were calculated to the points in transverse profiles at 0.01 m intervals (Figure 6 and Figure 7). DSM from TLS, as the most accurate measurement to the tacheometric surveying, was used as reference model to the other methods. The average width of transverse profile was 2.6 m and the profiles were interpolated after 0.01 m; which results approximately in 260 points per transverse profile. For the whole forest road, we are working with approximately 13,000 points (12,480 to be exact) in total.
5. As in the previous case, the DSMs produced by PLShh were compared with the most accurate DSM from TLS based on the simple difference of raster models and per pixel calculation. The compared digital surface models have a resolution of 0.01 m. There was a total of 4,304,481 points at which the evaluation was carried out.
The differences in accuracy for each comparison were again evaluated in MS Excel software. The deviations were then statistically evaluated. Positional and height differences between points of reference profiles and surface (from Faro Focus 3D) were calculated and the values for maximum, minimum, average, standard deviation, and RMSE (root mean square error) for each section and each device were calculated in the end altogether.

3. Results

When evaluating the accuracy of the laser scanning devices at the reference crosses, the positional and height deviations identified in the X, Y, and Z coordinates as well as the deviations of the lateral displacement in the X and Y axes were first evaluated. Table 2 shows very small mean error, in the order of less than 1 mm, which is due to the transformation of reference crosses to crosses targeted by the total station in the Cloud Compare software (Table 2). The mean deviation thus directly expresses the accuracy of the transformation. In the case of standard deviation, the data fluctuation is visible due to the ability of the device to capture the road surface. A comparison with the total station shows that the Faro Focus 3D scanner achieved the smallest positional deviation in XY with RMSE around 0.05 m and with smallest height deviation with RMSE around 0.01 m on all measured road sections. The GeoSLAM ZEB Horizon scanner achieved higher deviations (XY RMSE around 0.1 m and Z RMSE around 0.02 m) and the highest deviations were detected in the case of the iPhone 13 Pro with the 3D Scanner app and especially with the Polycam app (XY RMSE around 0.20 m, Z RMSE around 0.03 m). These results show that all scanning methods used are applicable for mapping directional and elevation changes of the road route.
In the next step, the ability of used laser scanning methods for length detection between the reference crosses in the direction of the route measured tacheometrically was assessed on each section on both sides. The error in length determination increased with distance for most of the laser scanning devices used. At a distance of 150 m, the error in length determination for the Faro Focus 3D scanner varied from 0.02 to 0.15 m, for GeoSLAM ZEB Horizon from 0.02 to 0.25 m, for iPhone 3D Scanner App from 0.53 to 0.98 m, and for iPhone Polycam App from 0.34 to 1.49 m (Figure 5) compared to tacheometric surveying.
When evaluating the length comparison to tacheometric surveying between reference crosses within the transverse profiles, these errors were slightly smaller from 0.02 to 0.09 m (Table 3).
As the data from the Faro Focus 3D scanner were found to have minimal positional and height deviations compare to tacheometric surveying (RMSE XY less than 0.05 m and RMSE Z 0.034 m) based on a previous evaluation of the results, these data were used as a reference for the subsequent evaluation of the transverse profiles and digital surface models height differences accuracy (Table 4 and Table 5). A graphical representation of the evaluation of the transverse profiles, using Stack Profile tool to calculate the heights from DSMs to the points in transverse profiles, is shown in Figure 6 and Figure 7. The results of both evaluation methods were very similar, with the GeoSLAM ZEB Horizon scanner achieving an RMSE of around 0.03 m, the Polycam app around 0.04 m, and the 3D Scanner app around 0.02 m.

4. Discussion

Geodetic methods for tacheometric surveying are used these days mostly for geodetic survey, and tacheometric surveying is therefore used in this article as a reference status when assessing the accuracy of used laser scanning methods. This study shows that TLS can replace tacheometric surveying with very similar accuracy. The advantage is that information about the road surface condition, including possible damage in the form of ruts, potholes, and other deformations is obtained. Nevertheless, even this method of data collection is time-consuming and therefore does not bring an advantage over current methods. At present, we can observe a great shift in the way data are acquired, as tacheometric surveying is being replaced by LiDAR technologies. However, these mainly concern roads in open countryside and urban environments [27,28], where a mobile laser scanner is placed on a motor vehicle. This technology was investigated in the forest environment by [19] as MLS, but their results show an error of the RMSE with vertical differences of 0.4228 m compared to scanning outside the forest, where the vertical differences of the RMSE reached only 0.0570 m. Such an error eventually excludes the use of MLS technology from practical use in the forest environment. Additionally, the information obtained so far mainly comprises road markings and their technical equipment, rather than the condition of the road wearing course. Still, there are already some advances in this direction as well, as some studies mention possible detection of damage size and depth of cracks in an asphalt surface [25,26].
Further development of technologies (especially the possible automatic scan registration, e.g., using the SLAM algorithm) leads to the use of smaller mobile devices, such as the GeoSLAM ZEB Horizon scanner, which are not dependent on the GNSS signal and allow for fast and continuous data collection. However, these devices are very expensive and require additional software for data processing. In addition, the study shows that the accuracy of the scan and the resulting point cloud is lower than that of TLS. A ground-breaking novelty is the use of LiDAR technology in mobile phones, currently only available in the iPhone 12, 13, and 14 Pro. Previous studies have demonstrated their ability to accurately capture objects, from individual trees in forest stands [35] to rock cliffs [20]. Our study shows results on forest road in forest environment. Surprisingly, if the 3D Scanner app is used, the results were even better than those of the GeoSLAM Horizon scanner. This may be due to the more difficult identification of reference crosses, as the collection did not use RGB camera point cloud staining and crosses were only identified based on the change in LiDAR reflection intensity, so identifying the centers of the reference crosses in the point cloud was sometimes difficult and less accurate. The comparison of the reference crosses position with the coordinates from the total station showed that all the methods used were able to capture the height profile of the road very well with deviations in the order of centimeters. However, in the case of the length measurement, LiDAR on the iPhone 13 Pro returned a length error of up to tens of centimeters.
The comparison of height deviations created by digital surface models of the surface shows that all instruments and applications were comparable in their accuracy. The best results were surprisingly achieved by the 3D Scanner app with an RMSE with about 2 cm. However, the results were also influenced by the positional deviation of the models given by their referencing to crosses. The graphical representation of the transverse profile (Figure 6 and Figure 7) shows that the GeoSLAM Horizon scanner was better able to capture local unevenness (such as road damage), while the applications on the iPhone 13 Pro smoothed the shape of the surface.
Data collection using the iPhone 13 Pro is fast and comparable in time to HMLS, and the price of the device is several times lower than the GeoSLAM Horizon scanner (1000 EUR vs. 40,000 EUR).

5. Conclusions

The aim of the study was mainly to determine whether it is possible to use the iPhone 13 Pro with LiDAR technology and the currently developed laser scanning applications 3D Scanner and Polycam app to determine the extent of damage on forest road surfaces in terms of their accuracy compared to in these days used tacheometric surveying or a static laser scanner. For further extension regarding other types of hand-held personal laser scanners, a ZEB HORIZON scan has been added.
Although the LiDAR in the iPhone 13 Pro still does not reach the quality and accuracy for professional use in the continuous acquisition of data along the length of the route, considering the price and speed of data collection, it is a very promising technology that allows fast data collection and, above all, easy mapping and inspection of forest roads without depending on the GNSS signal. The comparison of the length data collection show that the error increased with the length of the measured section, since the distance deviation in the transverse profiles was much lower in both iPhone 13 Pro apps than in the length measurement in the direction of the route in all sections.
The results show that the LiDAR in iPhone 13 Pro can capture the surface of forest road with sufficient accuracy, yet still only locally (in the cross section of the forest road up to 4 m). Progress regarding the length accuracy must be made and this should be verified by further research on all of tested PLShh, especially with both apps on the iPhone 13 Pro.). All PLShh are able to capture the current shape of forest road cross profile with given RMSE and the 3D Scanner app is able to provide information about changes with an accuracy of two centimeters. Moreover, as can be seen from the results, 3D Scanner app can capture the shape in the same progression as TLS as well. This can help forest managers with decision processes to determinate the urgency and set the order of forest road repairs on their estate in consideration of objective data and not based on personal decision. However, this cannot be still used for any design or calculation of material measurement statements needed for wearing course repair. Here, it must be emphasized as well that the analysis was carried out on a sealed wearing course and the given results should be related to this type of surface.
The advantage of both apps on the iPhone 13 Pro is the direct export of the created point cloud and possible mesh to various formats, which are further processable in most GIS and CAD applications. The practical usage is determined by the precision needs of a specific design.

Author Contributions

T.M. initiated the study and the methodology, discussion, carried out and reviewed the statistical analysis and drafted the manuscript. D.K. conducted data processing and the GIS and statistical analysis. M.C. and Z.P. carried terrestrial laser scanning and personal laser scanning and process the data. P.H. prepared the abstract, introduction, discussion, conclusion, contributed to the methodology and revised the text and the manuscript. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Internal Grant Agency of the Faculty of Forestry and Wood Technology, Mendel University in Brno, Czech Republic, grant number IGA-LDF-22-IP-027 “Possibilities of using a mobile laser scanning using a mobile device (smartphone) with integrated LiDAR for forest roads mapping”.

Data Availability Statement

Data contained within the article are available on request from the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Toscani, P.; Sekot, W.; Holzleitner, F. Forest Roads from the Perspective of Managerial Accounting—Empirical Evidence from Austria. Forests 2020, 11, 378. [Google Scholar] [CrossRef] [Green Version]
  2. The Parliament of the Czech Republic. Road Act and Amendments to Certain Acts (Road Act); Act No. 13/1997 Coll; on Forests and on Amendments to Certain Acts (Road Act); The Parliament of the Czech Republic: Prague, Czech Republic, 1997. [Google Scholar]
  3. The Parliament of the Czech Republic. Forest Act and Amendments to Certain Acts (Forest Act); Act No. 289/1995 Coll; on Forests and on Amendments to Certain Acts (Forest Act); The Parliament of the Czech Republic: Prague, Czech Republic, 1995. [Google Scholar]
  4. Dodson, E.M. Challenges in Forest Road Maintenance in North America. Croat. J. For. Eng. 2021, 42, 107–116. [Google Scholar] [CrossRef]
  5. Akgul, M.; Yurtseven, H.; Akburak, S.; Demir, M.; Cigizoglu, H.K.; Öztürk, T.; Eksi, M.; Akay, A.O. Short term monitoring of forest road pavement degradation using terrestrial laser scanning. Measurement 2017, 103, 283–293. [Google Scholar] [CrossRef]
  6. Wang, H. Development of Laser System to Measure Pavement Rutting. Master’s Thesis, University of South Florida, Tampa, FL, USA, 2005. Available online: http://digital.lib.usf.edu/SFS0025691/00001 (accessed on 7 November 2017).
  7. Demir, M.; Hasdemir, M. Functional planning criterion of forest road network system according to recent forestry development and suggestion in Turkey. Am. J. Environ. Sci. 2005, 1, 22–28. [Google Scholar] [CrossRef] [Green Version]
  8. Starke, M.; Kunneke, A.; Ziesak, M. Monitoring of Carriageway Cross Section Profiles on Forest Roads: Assessment of an Ultrasound Data Based Road Scanner with TLS Data Reference. Forests 2021, 12, 1191. [Google Scholar] [CrossRef]
  9. Yurtseven, H.; Akgul, M.; Akay, A.O.; Akburak, S.; Cigizoglu, H.K.; Demir, M.; Eksi, T.O.A.M. High accuracy monitoring system to estimate forest road surface degradation on horizontal curves. Environ. Monit. Assess. 2019, 191, 32. [Google Scholar] [CrossRef]
  10. Bhatt, A.; Bharadwaj, S.; Sharma, V.B.; Biswas, R.D.A.S. An Overview of Road Health Monitoring System for Rigid Pavement By Terrestrial Laser Scanner. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2022, 43, 173–180. [Google Scholar] [CrossRef]
  11. Fidelus-Orzechowska, J.; Strzyżowski, D.; Wrońska-Wałach, J.C.A.D. A Quantitative Analysis of Surface Changes on an Abandoned Forest Road in the Lejowa Valley (Tatra Mountains, Poland). Remote Sens. 2020, 12, 3467. [Google Scholar] [CrossRef]
  12. Hrůza, P.; Mikita, T.; Tyagur, N.; Krejza, Z.; Cibulka, M.; Procházková, A.; Patocka, Z. Detecting Forest Road Wearing Course Damage Using Different Methods of Remote Sensing. Remote Sens. 2018, 10, 492. [Google Scholar] [CrossRef] [Green Version]
  13. Hyyppä, J.; Hyyppä, H.; Yu, X.; Kaartinen, H.; Kukko, A.; Holopainen, M. Forest Inventory Using Small-Footprint Airborne LiDAR. In Topographic Laser Ranging and Scanning, Principles and Processing; Shan, J., Toth, C.K., Eds.; CRC Press: Boca Raton, FL, USA, 2009. [Google Scholar] [CrossRef]
  14. Brenner, C. Building Reconstruction from Images and Laser Scanning. Int. J. Appl. Earth Obs. Geoinf. 2005, 6, 187–198. [Google Scholar] [CrossRef]
  15. Sithole, G.; Vosselman, G. Experimental Comparison of Filter Algorithms for Bare-Earth Extraction from Airborne Laser Scanning Point Clouds. ISPRS J. Photogramm. Remote Sens. 2004, 59, 85–101. [Google Scholar] [CrossRef]
  16. Hyyppä, J.; Hyyppä, H.; Litkey, P.; Yu, X.; Haggrén, H.; Rönnholm, P.; Pyysalo, U.; Pitkänen, J.; Maltamo, M. Algorithms and Methods of Airborne Laser Scanning for Forest Measurements; International Society for Photogrammetry and Remote Sensing (ISPRS): Freiburg, Germany, 2004; Volume XXXVI–8/W2, pp. 82–89. [Google Scholar]
  17. Rodriguez-Perez, J.R.; Alvarez, M.F.; Sanz-Ablanedo, E. Assessment of low-cost receiver accuracy and precision in forest environments. J. Surv. Eng. 2007, 133, 159–167. [Google Scholar] [CrossRef] [Green Version]
  18. Bruggisser, M.; Hollaus, M.; Kükenbrink, D.; Pfeifer, N. Comparison of forest structure metrics derived from UAV lidar and ALS data. In ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences; Copernicus GmbH: Göttingen, Germany, 2019; Volume 4, pp. 325–332. [Google Scholar]
  19. Hrůza, P.; Mikita, T.; Janata, P. Monitoring of forest hauling roads wearing course damage using unmanned aerial systems. Acta Univ. Agric. Silvic. Mendel. Brun. 2016, 64, 1537–1546. [Google Scholar] [CrossRef] [Green Version]
  20. Luetzenburg, G.; Kroon, A.; Bjørk, A.A. Evaluation of the Apple iPhone 12 Pro LiDAR for an Application in Geosciences. Sci. Rep. 2021, 11, 22221. [Google Scholar] [CrossRef] [PubMed]
  21. Mrics, S.W.; Clegg, P.; Jones, R. Combining terrestrial laser scanning, RTK GPS and 3D visualization: Application of optical 3D measurement in geological exploration. In Proceedings of the 7th Conference on Optical 3D Measurement Techniques, Vienna, Austria, 3–5 October 2005. [Google Scholar]
  22. Barbarella, M.; De Blasiis, M.R.; Fiani, M.; Santoni, M. A LiDAR application for the study of taxiway surface evenness and slope. In ISPRS Annals Photogrammetry, Remote Sensing and Spatial Information Sciences, Proceedings of the ISPRS Technical Commision V Symposium, Riva del Garda, Italy, 23–25 June 2014; ISPRS: Göttingen, Germanny, 2014; Volume II-5. [Google Scholar] [CrossRef] [Green Version]
  23. Chin, A.; Olsen, M.J. Evaluation of technologies for road profile capture, analysis, and evaluation. J. Surv. Eng. 2015, 141, 04014011. [Google Scholar] [CrossRef]
  24. Jaakkola, A.; Hyyppä, J.; Hyyppä, H.; Kukko, A. Retrieval algorithms for road surface modelling using laser-based mobile mapping. Sensors 2008, 8, 5238–5249. [Google Scholar] [CrossRef] [Green Version]
  25. Guan, H.; Li, J.; Cao, S.; Yu, Y. Use of mobile LiDAR in road information inventory: A review. Int. J. Image Data Fusion 2016, 7, 219–242. [Google Scholar] [CrossRef]
  26. Kumar, P.; Angelats, E. An automated road roughness detection from mobile laser scanning data. ISPRS Arch. 2017, 42, 91–96. [Google Scholar] [CrossRef]
  27. Yan, L.; Li, Z.; Liu, H.; Tan, J.; Zhao, S.; Chen, C. Detection and classification of pole-like road objects from mobile LiDAR data in motorway environment. Opt. Laser Technol. 2017, 97, 272–283. [Google Scholar] [CrossRef]
  28. Vallet, B.; Papelard, J.P. Road orthophoto/DTM generation from mobile laser csanning. ISPRS Ann. Photogramm. Remote Sens. Spat. Inf. Sci. 2015, 2, 377–384. [Google Scholar] [CrossRef] [Green Version]
  29. Kukko, A.; Kaartinen, H.; Hyyppä, J.; Chen, Y. Multiplatform approach to mobile laser scanning. In Proceedings of the 2012 XXII ISPRS Congress, Melbourne, Australia, 25 August–1 September 2012; Volume XXXIX-B2, pp. 483–488. [Google Scholar]
  30. Kukko, A.; Kaijaluoto, R.; Kaartinen, H.; Jaakkola, A.; Hyyppä, J. Graph SLAM correction for single scanner MLS forest data under boreal forest canopy. ISPRS J. Photogramm. Remote Sens. 2017, 132, 199–209. [Google Scholar] [CrossRef]
  31. Gollob, C.; Ritter, T.; Nothdurft, A. Forest Inventory with Long Range and High-Speed Personal Laser Scanning (PLS) and Simultaneous Localization and Mapping (SLAM) Technology. Remote Sens. 2020, 12, 1509. [Google Scholar] [CrossRef]
  32. Balenović, I.; Liang, X.; Jurjević, L.; Hyyppä, J.; Seletković, A.; Kukko, A. Hand-Held Personal Laser Scanning: Current Status and Perspectives for Forest Inventory Application. Croat. J. For. Eng. 2021, 42, 165–183. [Google Scholar] [CrossRef]
  33. Chen, S.; Liu, H.; Feng, Z.; Shen, C.; Chen, P. Applicability of personal laser scanning in forestry inventory. PLoS ONE 2019, 14, e0211392. [Google Scholar] [CrossRef] [PubMed]
  34. Gollob, C.; Ritter, T.; Kraßnitzer, R.; Tockner, A.; Nothdurft, A. Measurement of Forest Inventory Parameters with Apple iPad Pro and Integrated LiDAR Technology. Remote Sens. 2021, 13, 3129. [Google Scholar] [CrossRef]
  35. Mokroš, M.; Mikita, T.; Singh, A.; Tomaštík, J.; Chudá, J.; Wężyk, P.; Kuželka, K.; Surovỳ, P.; Klimánek, M.; Zięba-Kulawik, K.; et al. Novel low-cost mobile mapping systems for forest inventories as terrestrial laser scanning alternatives. Int. J. Appl. Earth Obs. Geoinf. 2021, 104, 102512. [Google Scholar] [CrossRef]
Forests 13 01763 g001 550
Figure 1. The study area—forest road “Skatulova” located in the Training Forest Enterprise “Masaryk Foresk” Křtiny.
Figure 1. The study area—forest road “Skatulova” located in the Training Forest Enterprise “Masaryk Foresk” Křtiny.
Forests 13 01763 g001
Forests 13 01763 g002 550
Figure 2. Laser scanning instruments used for forest road mapping: (a) Faro Scene 3D (Faro Technologies Ltd., Rugby, UK), (b) GeoSLAM ZEB Horizon (GeoSLAM Ltd., Nottingham, UK), (c) iPhone 13 Pro (Apple Inc., Cupertino, CA, USA).
Figure 2. Laser scanning instruments used for forest road mapping: (a) Faro Scene 3D (Faro Technologies Ltd., Rugby, UK), (b) GeoSLAM ZEB Horizon (GeoSLAM Ltd., Nottingham, UK), (c) iPhone 13 Pro (Apple Inc., Cupertino, CA, USA).
Forests 13 01763 g002
Forests 13 01763 g003 550
Figure 3. Digital surface model of each scanning method visualised from a part of road section No. 3 with a fivefold elevation change.
Figure 3. Digital surface model of each scanning method visualised from a part of road section No. 3 with a fivefold elevation change.
Forests 13 01763 g003
Forests 13 01763 g004 550
Figure 4. Measurement of lengths in the direction of the path between reference crosses (red lines) and in transverse profiles (blue lines)—example from section No.1.
Figure 4. Measurement of lengths in the direction of the path between reference crosses (red lines) and in transverse profiles (blue lines)—example from section No.1.
Forests 13 01763 g004
Forests 13 01763 g005 550
Figure 5. Comparison of lengths in the direction of the route (example from right side of the section No. 2) to tacheometric surveying.
Figure 5. Comparison of lengths in the direction of the route (example from right side of the section No. 2) to tacheometric surveying.
Forests 13 01763 g005
Forests 13 01763 g006 550
Figure 6. Cross profile example—road section No. 1.
Figure 6. Cross profile example—road section No. 1.
Forests 13 01763 g006
Forests 13 01763 g007 550
Figure 7. Cross profile example—road section No. 3.
Figure 7. Cross profile example—road section No. 3.
Forests 13 01763 g007
Table
Table 1. Comparison of time required for data collection of all sections 450 m long of the forest road.
Table 1. Comparison of time required for data collection of all sections 450 m long of the forest road.
Scanning DeviceTotal Time in Minutes
Faro Focus 3D170
Geoslam ZEB Horizon20
iPhone Polycam30
iPhone 3D Scanner30
Table
Table 2. Statistical evaluation of positional and height deviations compared to the reference crosses from tacheometric surveying.
Table 2. Statistical evaluation of positional and height deviations compared to the reference crosses from tacheometric surveying.
FARO FOCUS 3D
MeanStd. Dev.RMSE
devX−0.000010.034690.03469
devY−0.000020.035630.03563
devZ−0.000020.007790.00779
devXY−0.046600.017370.04973
Geoslam ZEB Horizon
MeanStd. Dev.RMSE
devX0.000000.068350.06835
devY0.000020.084660.08466
devZ0.000000.025740.02574
devXY0.089760.061500.10881
iPhone Polycam
MeanStd. Dev.RMSE
devX0.000000.117720.11772
devY−0.000010.286780.28678
devZ−0.000010.045610.04561
devXY0.255680.175300.31000
iPhone 3D Scanner
MeanStd. Dev.RMSE
devX−0.000010.106090.10609
devY−0.000010.218950.21895
devZ0.000000.021900.02190
devXY0.164140.086010.18531
Table
Table 3. Statistical evaluation of accuracy—lengths between reference crosses in transverse profiles.
Table 3. Statistical evaluation of accuracy—lengths between reference crosses in transverse profiles.
Scanning DeviceMeanStd. Dev.RMSE
Faro Focus 3D0.010.020.02
Geoslam ZEB Horizon0.010.080.08
iPhone Polycam0.060.090.11
iPhone 3D Scanner0.050.050.07
Table
Table 4. Statistical evaluation of accuracy—height differences of points from transverse profiles in comparison with TLS scanner Faro Focus 3D.
Table 4. Statistical evaluation of accuracy—height differences of points from transverse profiles in comparison with TLS scanner Faro Focus 3D.
Scanning DeviceMeanMinMaxStd. Dev.RMSE
Geoslam ZEB Horizon0.0143−0.07740.10120.02880.0321
iPhone Polycam0.0082−0.11550.10930.04040.0413
iPhone 3D Scanner0.0067−0.05550.06030.01610.0174
Table
Table 5. Statistical evaluation of differences between digital surface models in comparison with reference model from TLS scanner Faro Focus 3D.
Table 5. Statistical evaluation of differences between digital surface models in comparison with reference model from TLS scanner Faro Focus 3D.
Scanning DeviceMeanMinMaxStd. Dev.RMSE
Geoslam ZEB Horizon0.008−0.2620.8100.0180.028
iPhone Polycam0.001−0.1200.8850.040.041
iPhone 3D Scanner0.008−0.1680.8320.0160.018
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Mikita, T.; Krausková, D.; Hrůza, P.; Cibulka, M.; Patočka, Z. Forest Road Wearing Course Damage Assessment Possibilities with Different Types of Laser Scanning Methods including New iPhone LiDAR Scanning Apps. Forests 2022, 13, 1763. https://doi.org/10.3390/f13111763

AMA Style

Mikita T, Krausková D, Hrůza P, Cibulka M, Patočka Z. Forest Road Wearing Course Damage Assessment Possibilities with Different Types of Laser Scanning Methods including New iPhone LiDAR Scanning Apps. Forests. 2022; 13(11):1763. https://doi.org/10.3390/f13111763

Chicago/Turabian Style

Mikita, Tomáš, Dominika Krausková, Petr Hrůza, Miloš Cibulka, and Zdeněk Patočka. 2022. "Forest Road Wearing Course Damage Assessment Possibilities with Different Types of Laser Scanning Methods including New iPhone LiDAR Scanning Apps" Forests 13, no. 11: 1763. https://doi.org/10.3390/f13111763

APA Style

Mikita, T., Krausková, D., Hrůza, P., Cibulka, M., & Patočka, Z. (2022). Forest Road Wearing Course Damage Assessment Possibilities with Different Types of Laser Scanning Methods including New iPhone LiDAR Scanning Apps. Forests, 13(11), 1763. https://doi.org/10.3390/f13111763

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop