Sustainable Ecotourism through Cutting-Edge Technologies

Abstract

Tourism is a phenomenon that dates back to ancient times. Ancient Greek philosophers recognised, adopted, and promoted the concept of rest-based tourism. Ecotourism is a particular type of tourism that connects with activities that take place in nature, without harming it, along with the herbal and animal wealth. According to estimates, the global ecotourism industry is currently booming due to various reasons, and it is becoming an important factor of sustainable regional development. This article presents the vision, work, and outcomes of project AdVENt, a project focusing natively in sustainable ecotourism through natural science and technological innovation. AdVENt’s study area includes the National Parks of Oiti (or Oeta) and Parnassus in Central Greece, where there is a remarkable native flora with a high endemism rate integrated with areas of cultural value and national and European hiking routes and paths of varying difficulty.

1. Introduction

Tourism is a phenomenon that dates back to ancient times. Ancient Greek philosophers recognised, adopted, and promoted the concept of rest based on tourism. During the Dark Ages (since 500 AD) tourism took the form of festivals and event participation. An aristocratic form of tourism appeared in the 16th century. The privileged classes were transferred to important destinations for the pursuit of both culture and education. Until the 18th century, emerging middle classes formed an ever-evolving element of tourism. In the early 1970s, tourism had begun using the natural and cultural resources of a destination as a tourist attraction [1,2,3,4,5].

Ecotourism is a form of alternative tourism (the term alternative tourism tends to diminish, nowadays, as the various traditionally referenced alternative forms become mainstream), in which activities take place in nature, without harming it, while also preserving the herbal and animal wealth. Ecotourism originates in ecological tourism. According to The International Ecotourism Society (TIES), ecotourism can be defined as

Responsible travel to natural areas that conserves the environment, socially and economically sustains the well-being of local people, and creates knowledge and understanding through interpretation and education of all involved (including staff, travellers, and community residents).

According to data presented at Statista (Statista, Market size of the ecotourism sector worldwide in 2019, with a forecast for 2027, online at https://www.statista.com/statistics/1221034/ecotourism-market-size-global/, accessed on 20 December 2021), the global ecotourism industry was estimated at USD 180 billion in 2019 with a prediction to climb to USD 330 billion in 2027 ( Figure 1). Ecotourism is booming due to various reasons, including

• Better infrastructures and access to ecotourism destinations.
• A pattern change from mass to personalised tourism.
• Increased interest in the protection of the environment.
• A shift in the perceived profile of an ecotourist, as different and eco-friendly
• Citizens’ engagement in the protection of the environment.
• Potential contributions to local economies.

National parks are natural ecosystems with spectacular ecological and scientific value but also with a rich aesthetic and cultural significance. National parks are important for their potential in biodiversity conservation and for their appeal to visitors and tourism, particularly ecotourism. The Mediterranean Basin is one of the world’s 18 important biodiversity regions, with a high geomorphological and climatic heterogeneity [6]. Its flora includes approximately 25,000 plant species or nearly 30,000 species and subspecies [7]; this makes it the richest among the Mediterranean ecosystems. In addition, Mediterranean mountains are particularly known for their vast plant diversity and endemism, ideal for biogeographical and biodiversity studies. Greece is an important region of the Mediterranean Basin, being the actual meeting point of three continents, Europe, Asia, and Africa [8,9,10,11], characterised by its Mediterranean climate with mild, humid winters and hot, dry summers. Despite the general Mediterranean character, the diverse topology of the country results in differences in the regional natural environment and climate [12]. Consequently, Greece includes a large variety of Mediterranean habitats that have been referenced by the Natura 2000 initiative. Furthermore, the vascular flora of Greece includes 5758 species and 1970 subspecies, both native and naturalised, representing 6620 taxa, belonging to 1073 genera and 185 families. The endemic vascular flora of Greece includes 1459 taxa (22% of the total number of taxa in Greece), which correspond to 1274 endemic species (22.1% of the total number of Greek species), and 450 endemic subspecies (22.8% of the total number of Greek subspecies) [13,14]. The Greek National Parks are among the protected areas in Greece, which host a spectacular biodiversity and present a significant amount of ecological, aesthetic, cultural, educational, and scientific value. They offer a significant potential for conservation and attract a large number of visitors.

On the front of ecotourism innovation, there is a large volume of published research. Focusing only on relatively recent works, in 2016, Dionisio et al. presented a framework for trans-media storytelling in ecotourism, incorporating notions from location-based services, immersive multi-layered experiences, storytelling techniques, and mobile and social media technology [15]. The work focused on the presentation of the world-known natural wealth of Madeira Island. In 2019, Cranmer, although not specifically focusing on ecotourism, presented a study to identify key features for the effective design of Augmented Reality (AR) tourism applications [16]. This study identified four (4) AR design factors to be taken into account, namely, the visitor (knowledge transfer), the organisational (competition and sustainability), the stakeholder (preservation of knowledge), and the economic factor (regional sustainability), and set a preliminary framework of principles and guidelines for AR in tourism. Focusing on cultural tourism, [17] studied the increasing trend towards Virtual Reality (VR) applications. Of particular interest in this study was a fast growing group of users, the elderly; thus, it is also relevant for any tourism application. The results of this study were categorised under two major categories, analysing (a) the impact of VR on the experience and (b) the particular requirements for using VR. On category (a), the study reported a positive impact, and on (b), storytelling, presence, immersion, and control emerged as highly important aspects. A study that suggested a further distinction among users was presented by Armingeon et al. in 2019 [18], adopting the notions of digital natives and immigrants (to denote the generations after and before 1980, respectively) and trying to identify differences in the AR experiences in tourism. Interestingly, the survey data highlighted that digital natives rated their experience slightly worse, due to having higher expectations of AR than digital immigrants. This is an important finding for the future deployment of AR tourism experiences. Han et al. focused on the visitors’ learning experience in VR- and AR-supported tourism services [19]. The researchers recognised the key experience affecting elements, such as attention, involvement, immersion, cognitive absorption, and emotion, and highlighted that technology, and particularly AR and VR, should support the concentration and motivation of learners. Furthermore, they proposed a new conceptual framework for VR/AR-application impact assessment in cultural tourism. A market-oriented study was presented by Moorhouse et al. in 2019, focusing on providing insights into tourism marketers’ perspectives on adopting AR for enhanced tourism experiences [20]. This work identified opportunities and challenges for the adoption of AR technology from the marketers’ and the tourists perspectives, highlighting, as well, organisational, environmental, and infrastructure challenges. A systematic review that investigated the emerging information technology applications in sustainable tourism using the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) approach was presented in [21]. This review, which included published works since 1999, resulted in interesting statistics on the focus of research and highlighted the most frequently adopted technologies in sustainable tourism research. Liujin and Hong, in 2021, discussed the application of VR in an ecotourism exhibition [22]. They highlighted the challenges and presented a complete approach towards the development of large-scale VR applications based on modern game engines and Web3D technologies.

Project AdVENt (AdVENt—Augmented Visitor Experience in the National Parks of Mounts Oiti and Parnassus, http://advent.indigital.gr, accessed on 20 December 2021), a Greek national project in sustainable ecotourism, targets the development of innovative applications for the augmented experience of visitors in areas of significant natural and cultural value. The project draws on the natural and cultural value, the significant hiking trails, the picturesque traditional villages, the dense fir forests, and the endemic and rare herbal species of two of the most important National Parks of Greece. This study presents an overview of the innovations in AdVENt that bring new tools and possibilities towards sustainable regional development.

2. Objectives of Project AdVENt

Project AdVENt covers specific mountainous areas of the geographic region of Central Greece (Sterea Ellada), i.e., the National Park of Oiti and Parnassus, as shown in Figure 2. These parks provide the visitors with a variety of natural and cultural “gems” and a wide range of outdoor activities. There are hiking trails, dense fir forests to be explored, and endemic and rare plant species, all found in landscapes of unmatched natural beauty, not to mention famous historical and archaeological sites (Delphi) and popular winter sport destinations (Parnassus ski resort).

Mountain Oiti is also known as the mountain of flowers due to the amazing plethora of native and endemic flowers found there. It is also remembered as the place that marks the “ending” of the life of the mythical hero Hercules, who chose mountain Oiti for his funeral pyre (ruins of an ancient temple are found at the entrance of Katavothra Plateau). The unparalleled beauty of the mountain with its rare and impressive plant species, rich fauna, and abundant waters, flowing incessantly during all seasons, through steep and beautiful gorges, are markedly notable. The plateaus of Oiti, which cover about 10% of its area, and its rocky slopes, host the rarest and most interesting plant species, such as the local endemic flower Veronica oetaea and a species of wild onion, Allium lagarophyllum. Particularly impressive wildflowers are the narcissus of poets (Narcissus poeticus), which blooms at high altitudes in early summer, and the red lily (Lilium chalcedonicum). The forests of Oiti host a wide variety of fauna, of both common/widespread species, such as roe deer, wild boar, fox, hare, wildcat, etc., and rare and threatened species, such as the Balkan chamois (Rupicapra rupicapra ssp. balcanica), the brown bear (Ursus arctos), wolves (Canis lupus), as well as a wide variety of bird species. Mountain Oiti is the fifth highest mountain of central Greece, and its highest peak is Pyrgos (2152 m) (The Management Body of mountain Oiti National Park, https://oiti.gr/en/home/, accessed on 20 December 2021) [23].

Parnassus Mountain (Nature 2000 site code GR2410002, https://natura2000.eea.europa.eu/Natura2000/SDF.aspx?site=GR2410002, accessed on 20 December 2021) is located almost in the centre of Greece, spreads over three municipalities, and is considered as one of the largest mountainous regions of Mainland Greece. Its highest peak is Liakoura (2457 m), and Kelaria and Fterolakka are famous ski resorts. This mountain has spectacular geomorphological structures, steep cliffs, rocky areas, and caves. Its main element, limestone (76.6%), has helped in the creation of extraordinary formations, such as the sinkhole of Lilaia and the Corycian cave (sacred to the mythological god, Pan), which are famous for their wild beauty. Parnassus National Park and the National Park of Olympus are the two oldest national parks in Greece. On Parnassus one may find a high variety of flora, i.e., black pines, junipers, wild peonies, mountain tea, and rich fauna, namely, grey wolves, wild boars, ferrets, foxes, vultures, hawks, and reptiles. Parnassus used to be a sacred mountain and is also considered to be home to the nine Muses. The cultural and historical value of the place is great, and this is signified by the presence of the Sanctuary of Apollo and the Oracle of Delphi sites (The Management Body of Parnassus National Park, http://en.parnassosnp.gr, accessed on 20 December 2021).

The two selected sites belong to the Sub-Humid (SH) and Humid (H) climatic zones, respectively, according to the Climatic Zone Classification of UNEP (1992) [24], which is based on the Thornthwaite’s Aridity Index [25,26]. More specifically, in Oiti, the available data from the Agathonos Monastery meteorological station (38°52′2″ N, 22°12′16″ E, alt. 533 m), established by the Institute of Mediterranean Forest Ecosystems, covering the time period 1994–2004 indicate that the annual average air temperature is 13.7 °C, with seasonal variation between 5.5 °C in winter and 23.5 °C in summer, presenting intermediate values during the transitional seasons of spring and autumn (11.4 and 14.3 °C, respectively). The annual precipitation is 790.2 mm, occurring mainly in winter (336.1 mm) and spring (201.9 mm) and presenting diminished values in summer (59.0 mm). From the long-term available data of the nearby, though altitudinally lower, meteorological station of Lamia (38.9° N, 22.4° E, alt. 144 m), established by the Hellenic National Meteorological Service (HNMS), the values of the Aridity Index has been estimated since the beginning of the century and found a value of 0.64 (SH climate zone) for the latest climatic period (1960–1997), while the previous periods the Aridity Index values were increased, i.e., 0.75 for the climatic period 1930–1960 (H climate) and 0.69 for the climatic period 1900–1930 (H climate), indicating that the climate in the region became drier during the last decades [27]. In Parnassus, the nearest meteorological station was installed in Arachova (38.47° N, 22.57° E, alt. 950 m) by the HNMS. From the available data of the period 1976–1997, the annual average temperature in the region is 13.2 °C, seasonally ranging from 5.0 °C in winter to 22.3 °C in summer. The annual precipitation is 501 mm, which is unevenly distributed among seasons, occurring mainly in winter (191 mm) and autumn (150 mm) and less in spring (114 mm) and summer (45 mm). The Aridity Index value is 0.66 [27], corresponding to the H climate zone [24].

Project AdVENt set four main objectives:

(a)

To research on the vegetation of the region of interest and to create an online resource of the vascular plants of the Oiti and Parnassus National Parks available to the scientific community and the general public, aiming to promote research in the relevant fields of botany, ecology, etc.;

(b)

To digitise in 2D, 3D, and 360° selected natural and cultural treasures of the region for the creation of new rich digital content for the presentation and promotion of the region;

(c)

To develop an integrated Augmented Reality (AR) mobile application to enrich the tourist experience, particularly customised for the accommodation of mountaineering and hiking;

(d)

To research and develop AI technology and mobile applications for the image-based identification of plant species of the region of interest.

Overall, AdVENt aims at promoting and generally enhancing the visibility and competitiveness of the local tourism product, towards the support of regional sustainable development through innovative ecotourism. To achieve the AdVENt goals, a specific implementation methodology was designed and followed and was built on three major axes: (a) the research of the regional flora, (b) the digitisation and documentation, and (c) the research on AI approaches and mobile AR applications. According to this methodology, in the beginning, these three implementation directions were pursued independently and later converged to form a unified system. The following section describes the innovation in AdVENt on the basis of these three axes.

3. Innovation in AdVENt

Figure 3 summarises the innovations proposed in project AdVENt. A new scientific resource for regional flora and cutting-edge technologies are the main ingredients, empowering an ecotourism portal and an Augmented Reality (AR) mobile app, along with a scientific flora online resource. The following sections focus on the details of the key AdVENt innovations.

3.1. Resource of Vascular Plants of the Oiti and Parnassus National Parks

3.1.1. Methodology

A significant milestone in the implementation of AdVENt was the creation of the new regional flora resource. Development of a resource is a demanding process, requiring bibliographic and basic research, along with the creation of digital tools and infrastructure to support it. This was the case for this project. At the beginning, a wide bibliographic research took place, using offline and major online sources, including Google Scholar, Scopus, ScienceDirect, SciFinder, PubMed, along with the Forest Library of the Institute of Mediterranean Forest Ecosystems (IMFE) of the Hellenic Agricultural Organisation “DEMETER”, focusing on the hiking trails and the flora of the Oiti and Parnassus National Parks. The resource was planned to serve also as a source for relevant image content, so that a new flora dataset would be formed to support further development of Artificial Intelligence (AI) automatic-recognition methods. Thus, overall, the formation of the resource included the bibliographic research and on-site fieldwork at different times of the year (e.g., the flowering period) for the recording of species and the accumulation of multiple photographic records.

For the creation of the vascular plant database of Oiti and Parnassus, plants growing along the main pathways of these two National Parks were mainly targeted, augmented by the vast majority of plant species. Species-distribution data in the study areas were extracted from [28,29,30,31,32]. Plant nomenclature follows [14]. The database includes 832 plant taxa (species and subspecies), of which 676 taxa are distributed in Parnassus and 603 in Oiti, with an apparent overlap for common taxa in both mountains. The most diverse plant families are Asteraceae (85 taxa) and Fabaceae (74 taxa). The endemic floristic element is represented by 109 taxa, among which one may find several rare or threatened species, as well as seven (7) local endemics. The flora of these areas also includes a lot of socioeconomically important plants, like medicinal plants and crop wild relatives, highlighting the high value of the native flora and generally the significance of phytogenetic resources.

3.1.2. Implementation and Results

Collection of the data was a long-lasting process, including bibliographic and on-site research and digital recording. Initial digital records were kept in a cloud spreadsheet, for simplicity and remote multi-user access, consisting of a large number of fields, which served as potential candidates for the resulting data resource. Later, in a number of design iterations, the set of fields was finalised and the complete plant database structure was presented in [33], consisting of information for plants species as shown in

Table 1. The main structure of the plant species resource in AdVENt.

Field Name	Field Description
Group	The large group (i.e., Pteridophytes, Gymnosperms, or Angiosperms) to which the species belongs
Family	Family of the species
Genus	Genus of the species
Taxon	Species/subspecies
Author(s) citation	Authorship of the species
Scientific name	Valid scientific name of the species
Common name	Common name of the species
National extinction risk	National IUCN Red List threat category of the species
Global extinction risk	Global IUCN Red List threat category of the species
Status	Characterisation of the species as native, alien, and range restricted according to [13]
Endemic	Species with a distribution restricted to the territory of Greece
Crop wild relative	Indication that the species is a wild relative of a crop plant according to the definition provided by [34]
Medicinal	Indication that the species has a known medicinal use
Regional distribution	Distribution of the species in the 13 floristic regions of Greece according to [13]
Life-form categories	Life-form category of the species according to [13]
Habitat categories	Habitat categories of the species according to [13]
Flowering period	Starting and ending months of the flowering period of the species
Description	Description of the main morphological characters of the species
Other information	Additional available information for the species, e.g., traditional uses, historical data, mythology and art, etc.
References	The available references related to the species

.

In order to publish this set of valuable data and make it available as a scientific resource on the web, a dedicated web portal capable to support dynamic content management and API-based connectivity and data dissemination was developed. The initial cloud spreadsheet was transformed into a complete web database, ingesting all the scientific data. The integrated Database Management System (DBMS)—front-end—API system supports the online content management by the experts, the publishing of articles, the map-based or index-based search, and the retrieval of scientific flora data, as depicted in Figure 4. Figure 5 showcases the basic front-end interface with some statistics and the presentation form (the AdVENt scientific Greek vascular plants resource can be accessed online at http://advent.athenarc.gr, accessed on 20 December 2021).

3.2. Digitisation and Content Creation

3.2.1. Methodology

A factor of major importance in successful tourism applications is the quality of the content. An ecotourism application, such as the one AdVENt envisioned, bridging elements of natural and cultural value, on the founding element of hiking tourism, must integrate various experiences. The basis for the development of the content in AdVENt was the mixture of the selection of interesting and important hiking trails along with Points of Interest (POIs). Apparently, this is not an easy task; thus, criteria were established to make a selection possible, by investing a considerable amount of time and resources for on-site and online surveys.

The collection of online sources and information was intended to assess the trails, sites, and POIs applicable to the goals of the project. The selection criteria used can be summarised to (a) the attractiveness in terms of natural beauty, (b) the traffic, (c) the level of difficulty in the trails, (d) the accessibility to POIs, (e) natural wealth in terms of flora and characteristics of natural environment, and (f) points of cultural interest involving trails. Relevant online data were collected by contacting authorities, accessing open data and running web-crawling sessions, resulting in a wealth of published studies and texts on geomorphology, ecology, history, and culture of the areas of interest. A preliminary list of mountainous routes and trails of Oiti and Parnassus was created, including information on their basic characteristics, like the level of difficulty, length, altitude difference, route description, and relevant POIs.

Apparently, an online survey is not enough for a project of ecotourism; thus, on-site visits were planned and executed, to experience the parameters of each of the possible available selections. Further criteria were imposed, including (a) the need for each trail to combine natural and cultural value and (b) the need for each point of cultural interest to be clear of vegetation and obstacles in order to be accessible for digitisation. Additionally, the POIs should be popular and remarkable in terms of natural beauty and environmental interest, and the trails should be representative of the natural environment and morphology of the study areas.

3.2.2. Implementation and Results

The digitisation of trails and POIs in a mountainous area such as that of National Parks is a demanding process. The lack of accessibility by a vehicle at various points of a trail or a route, the lack of mains power throughout the process, the large altitude variations, and the long length of certain trails are definite threats to outdoors’ digitisation. Careful planning and selection resulted in a set of 30 hiking trails and another set of 70 connected POIs. The POIs were either of natural or cultural significance and were distributed throughout the study area as shown in Figure 6.

The data-acquisition process included the selection of multiple media for various applications. A general set of thousands of photographs per POI was collected for the application of ground and aerial photogrammetry and the 3D reconstruction of locations of natural and cultural value. Figure 7 showcases the 3D digitisation process with two examples, a religious monument and a medieval castle; on the left, the hill and remains of the Medieval Castle of Ypati on mount Oiti, in which around 1000 aerial photos were collected, resulting in a 3D reconstruction with a 1 pixel reprojection error for a ground resolution of less than 6 mm per pixel; on the right, there is the Hagia Sophia byzantine church on mount Oiti, which was recorded by around 1400 ground and 100 aerial photos that resulted in a 3D reconstruction of a 1 pixel reprojection error and a resolution of around 0.5 mm (0.25 mm per pixel).

Three-dimensional reconstruction examples of various scales are showcased in Figure 8. Monuments of varying size, caves, and even landscapes were reconstructed to create rich multimedia content to support the project applications. In addition, in POIs in which 3D digitisation could not be applied due to accessibility or coverage reasons, 3D modelling was used, as shown in Figure 9 for the cases of the museum of water power and the church of prophet Elias in Pavliani area. In addition to digitisation and modelling of POIs, 360° spherical panorama videos and photos were captured to offer an alternative, but still immersive, presentation paradigm, as illustrated in Figure 10. Last but not least, all selected hiking trails and paths were captured in first-person video, by a professional mountain guide, using action cameras and stabilisers along with spatial audio. Additional sounds of nature in POIs were also captured in order to enrich the apps and visualisations with the corresponding audio experience.

3.3. Cutting-Edge Technologies

3.3.1. Methodology

Artificial Intelligence (AI) and Augmenter Reality (AR) were selected as the cornerstones for the innovation in AdVENt. AI was used for the development of algorithms and systems for the automatic identification of plants in the study area, and AR was used to augment an ecotourist visit with rich and useful content to guide and enhance the mountaineering experience.

3.3.2. Implementation and Results

The research for the creation of the vascular-plants digital resource for the mountains of Oiti and Parnassus resulted in a new image dataset that was published under the name GRASP-125 [35]. The GRASP-125 dataset (The Greek Vascular Plants GRASP-125 dataset, http://advent.athenarc.gr/grasp/, accessed on 20 December 2021) was composed of 125 classes of different species of vascular plants. Within those species there are significant morphological variations depending on the developmental stage, the time of day, the season and even the location where each individual is growing. These variations can affect the morphological characteristics either of the whole plant or the appearance of its individual organs, such as flowers, leaves, and fruits at various stages in its life cycle; thus, any automatic identification of plant species requires special treatment in terms of robustness to deformations and scene flow [36,37,38,39].

As already known, contemporary AI image-recognition techniques require a heavy amount of training data. GRASP-125 consists of 125 species with an average of 130 images per class, which amounts to about 16,000 images. The dataset is open and intended to facilitate further research in the domain, which is in contrast to other proprietary datasets that were created in a controlled laboratory environment or other rather “noisy” crowd-sourced datasets. Till the publication of GRASP-125, the most widely used image datasets of plants consisted either of scans or pseudo-scans of leaves or images of different plant organs (leaves, flowers, stems, and fruits), and more complex photos (branches of trees or the entire plant), captured in their natural environment. There are at least four (4) popular leaf image sets, including the Swedish leaf (The Swedish leaf dataset, http://www.cvl.isy.liu.se/en/research/datasets/swedish-leaf/, accessed on 20 December 2021), the Flavia (Flavia plant leaf recognition system and dataset, https://sourceforge.net/projects/flavia/, accessed on 20 December 2021), the Leafsnap (The Leafsnap dataset, http://leafsnap.com/dataset/, accessed on 20 December 2021), and the ICL (The ICL dataset, http://english.iim.cas.cn, accessed on 20 December 2021). Other popular datasets provide flower images captured in the natural environment, like the Oxford Flower 17 and 102 (Flower Datasets, https://www.robots.ox.ac.uk/~vgg/data/flowers/, accessed on 20 December 2021) and the Jena Flower 30 (Jena Flowers 30 Dataset, https://doi.org/10.7910/DVN/QDHYST, accessed on 20 December 2021). The richest image dataset for plant identification is provided by the Image Cross Language Evaluation Forum (ImageCLEF), which has been organising a plant-identification challenge (PlantCLEF) each year since 2011, accumulating massive amounts of data. Another large dataset is the German flowering plants dataset [40]. The GRASP-125 dataset consists of two complementary sets of plant images collected as field photos from the study area and as web images that were verified by expert botanists. The life forms represented in the GRASP-125 dataset consist of trees, shrubs, and herbs growing at an altitude of over 1000 m on the mountains of Oiti and Parnassus. Figure 11 presents a small sample from the GRASP-125 dataset.

Based on GRASP-125, a large-scale experiment was setup to assess the value of the dataset in automated plant identification. Several popular Convolutional Neural Network (CNN) architectures were used for the task, using the transfer learning approach. [35] reported in detail on the performance of each of the tested architectures and on the explainability of the outcomes, using SmoothGrad [41] and Grad-CAM++ [42]. The attained performance results for the classification of the GRASP-125 dataset are summarised to a 92% top-1 and 98% top-5 accuracy. The analysis of all the architectures and their performances, along with constraints relating to lightweight mobile applications and web services, guided the final selection of the EfficientNetB4, coupled with transfer learning, to be the choice adopted for the plant identification in AdVENt (the plant identification git repository can be accessed online at https://gitlab.com/ilsp-xanthi-medialab/public/advent/advent-plant-recognition, accessed on 20 December 2021, whereas the web service for the plant identification can be tested online at http://services.athenarc.gr:5000/advent/gui, accessed on 20 December 2021). The validation loss and validation accuracy of the selected model, EfficientNetB4, fine-tuned with GRASP-125 is shown in Figure 12.

Furthermore, the explainability of the identification results for Greek vascular plants by EfficientNetB4 is shown in Figure 13, in terms of the Grad-CAM++ visualisation and the confusion matrix for the most challenging GRASP-125 classes. The confusion matrix shown in this figure is only partial, as the complete matrix consists of $125\times 125$ elements and cannot be displayed properly; the chosen part shows one of the most challenging classes in the dataset. Two examples of flower recognition are shown in Figure 14; the screenshots were captured during experimentation with the algorithms and the mobile app (later on this) on site; on the left, there is a correct identification with a high rate, whereas on the right, there is a Brassicaceae family member interpreted as an otherwise similar in shape and colour Hippocrepis emerus, with a relatively low identification rate.

Google’s Flutter (Flutter platform, https://flutter.dev, accessed on 20 December 2021) open-source software development kit was the basis upon which the mobile app for the augmented ecotourism was built. The selection was made particularly due to the cross-platform delivery capabilities. The Wikitude library (Wikitude AR, https://www.wikitude.com, accessed on 20 December 2021) was included to enable AR functionalities and readily support POI-based presentations. Figure 15 and Figure 16 present the sequence diagrams for the mobile app, including the main sequence, the sequence for the automatic plant identification, the sequence for the interaction with hiking trails, and the sequence for the interaction with POIs. Figure 17 shows two examples of the AR-based guidance provided by the mobile app on-site, regarding the relative location of POIs, including directions and content relating to those POIs.

The open-source WordPress content management software was used as the basis for the development of the ecotourism web portal of AdVENt. The use of this platform enables the parameterisation of dynamic Web pages to support database connectivity and dynamic presentation. The platform supports multiple registered users of diverse access levels. In the portal, a visitor is able to choose an area of interest and browse the map where the POIs appear, along with geographic milestones like recorded hiking trails and paths that can be followed. The choice of a POI results in a particular presentation of the rich multimedia content attached to the selected POI, including 2D, 3D, and 360° content. Figure 18 shows the basic block diagram for the structure of the Oiti and Parnassus ecotourism portal. Figure 19 shows a screenshot from the AdVENt ecotourism portal showing the interactive map of mount Oiti with all the recorded POIs and trails. Selection of a POI results in the presentation of the content relating to that point, as shown in Figure 20, which presents a very peculiar location on the mountain with the church of Hagia Ierousalim (or, as known, Arsali) inside a cave. In addition, Figure 21 shows an example of the interactive map presenting a hiking trail on Oiti (the AdVENt ecotourism portal can be accessed online at https://new-advent.indigital.gr, accessed on 20 December 2021).

4. AdVENt’s Contribution and Multiplier Effects

AdVENt is expected to have a significant impact on the scientific community, society, and the economy. This section summarises AdVENt’s multi-dimensional and multi-modal contribution. It is multi-dimensional, since it covers multiple aspects of ecotourism support, and it is multi-modal, since it brings both scientific innovation and sustainable development support.

As a result of the interdisciplinary research in AdVENt, the project promoted research into the scientific field of AI, particularly applied to the recognition of flora species, towards the development of the first plant-identification system in Greece. As a result, a new Greek vascular-plant image dataset was created to support further research in AI image-analysis applications, such as GRASP-125, and is available online at http://advent.athenarc.gr/grasp/, accessed on 20 December 2021. The AI image-based plant identification system created in AdVENt is also available as a git repository that be accessed online at https://gitlab.com/ilsp-xanthi-medialab/public/advent/advent-plant-recognition, accessed on 20 December 2021. In addition, a web service for plant identification was created to support independent researchers, as well as the other AdVENt applications, and is accessible online at http://services.athenarc.gr:5000/advent/gui, accessed on 20 December 2021.

A resource of Greek flora was produced for the Oiti and Parnassus National Parks, available to the research community and the general public, which is expected to promote research in the relevant knowledge fields (botany, ecology, etc.), given that only limited relevant research efforts to record Greek flora have resulted in organised and widely available resources. The git repository for the scientific Greek vascular plants resource is accessible online at https://gitlab.com/ilsp-xanthi-medialab/public/advent/advent-plant-resource, accessed on 20 December 2021. For those interested in browsing or searching the scientific Greek vascular plants resource using our front-end, they may access it online at http://advent.athenarc.gr, accessed on 20 December 2021.

The project also promoted applied research in the field of 3D/360° digitisation and virtual reconstruction of cultural and natural spaces of small to large scale, using advanced digital-image recording activities. The rich content is available through the mobile app and the ecotourism portal of AdVENt.

As described in the previous sections, the project resulted in innovative and rich-in-content technological applications suitable for the promotion of the environmental wealth, the natural beauty, and the significant cultural value of the selected National Parks, as a modern and attractive tourist product, with a direct positive impact and development prospects for the local but also the wider economy and society. Taking into account that tourists of alternative tourism avoid organised tourist packages and organise their visits themselves, basically exploiting the vast resources on the Web, the AdVENt ecotourism portal is envisioned as an instrument for dissemination and promotion of the selected National Parks and the wider region, enhancing their recognition and, consequently, increasing their competitiveness as tourist destinations in Greece and abroad. The AdVENt ecotourism portal can be accessed online at https://new-advent.indigital.gr, accessed on 20 December 2021. In addition, AdVENt enhances the local tourist product with innovative AI and AR technological applications that enrich the on-site ecotourism experience by coupling the rich natural wealth content with cultural information. This coupling is compatible with the perceived profile of ecotourists who emphasise the regional, societal, and cultural identity and the local economy. The enhanced experiences are expected to strengthen visitor satisfaction, to increase positive opinion and word-of-mouth promotion, and to increase the competitive advantage and the visitor flows in the regions. The packaging of this concept into a mobile app that supports augmented experiences and offers connections with the AI services was a cornerstone for AdVENt, and the Android version can be found on the Google Play store at https://play.google.com/store/apps/details?id=gr.indigital.advent, accessed on 20 December 2021.

AdVENt’s results directly support national strategies regarding the technological advancement of products and services in the domain of tourism. Apparently, AdVENt supports the Regional Operational Program of Central Greece (the geographical location of the selected National Parks), one of the strategic axes of which is the diversification of the tourism product, taking advantage of the natural and cultural wealth of the region. Given the planned, ongoing reinforcement and upgrade of the tourist infrastructure in the area, the project’s applications can further support the region’s visibility and, at the same time, promote local tourism enterprises.

Last but not least, of significant importance in AdVENt is the transfer of know-how from the research organisations in the AdVENt consortium to the partner SMEs, which, evidently, increases their competitiveness overall, and is a global best practice. For the three-year duration of the project, more than ten (10) new scientists were employed full-time, strengthening their professional experience and creating the conditions for expanding their cooperation with research entities and businesses in the future.

Regarding the future of AdVENt, there have been preliminary exploitation plans that are based on the technology innovation and modular development followed during implementation. Having all the infrastructure and basic apps developed, the results are easily transferable and applicable to other regions and countries. This publication serves as a dissemination instrument towards this direction. In addition, all the developed technologies and apps, although in prototype, can be easily brought to higher Technology Readiness Levels (TRLs), which denote how ready a product or service is to be commercially marketed, and can be commercially exploited, either by using advertisement, direct sales, or subscription models, depending on the target groups and the envisioned impact. The integrated enhanced ecotourism experience supported by AdVENt is among the first worldwide and certainly the first of its kind in Greece, representing the next step in the usual online and printed guides. This is per se a competitive advantage for the AdVENt partners in the domain of ecotourism. Currently, the partnership of AdVENt is designing a detailed marketing and exploitation plan, hoping that the innovations created in the project may bring significant benefits towards a sustainable ecotourism.

5. Conclusions

Ecotourism is becoming a highly dynamic type of tourism, with a fast-growing share in the tourism economy. Its successful future is based on multiple factors, particularly those that connect tourists with the natural and cultural environment. Thus, ecotourism is expected to support sustainable regional development. This article presented the multi-dimensional results of an ecotourism innovation project, AdVENt, which focused on sustainable ecotourism through natural science and technological innovation. AdVENt targeted its activities on two National Parks in Central Greece, which include natural and cultural wealth for mountaineering, hiking, and recreational and educational activities, while being of particular importance due to their diverse flora and landscapes. AdVENt applied basic research on the endemic vascular plant flora of the study area and applied cutting-edge ground and aerial 3D-digitisation approaches, 3D modelling and 360° recording technologies, AI approaches, and AR techniques to implement the vision of sustainable tourism through technological innovation, delivered through the web and smart mobile devices. The integrated final results of AdVENt include a scientific online resource for the Greek vascular plant flora, a web portal for the National Parks of Oiti and Parnassus offering multi-dimensional experiences, and a mobile app for the augmentation of an on-site visit with rich content and helpful guides.