Infrared Imaging System with a Local Polarization Channel for Target Detection

Abstract

Infrared imaging can detect the targets from their backgrounds during the day and night. As a supplementary tool, polarization imaging can visually characterize the polarized information of targets in complicated environments. In this study, an infrared imaging system with a local polarization channel is developed and simulated. The system can acquire local infrared and polarization information of the target without compromising the monitoring of the entire field of view. This system is composed of two infrared imaging channels with a common image plane: a peripheral channel for the entire field of view and a local imaging channel for the target of interest to achieve local magnification and polarization imaging. Herein, an imaging simulation of the infrared imaging system with a local polarization channel is performed, and the results show that the system performs well.

1. Introduction

Infrared imaging technology [1,2,3] perceives the difference in radiation energy between objects and their backgrounds; thus, it achieves target detection and identification. The demand for all types of infrared imaging devices [4,5,6] is continually growing in numerous fields, such as remote sensing [7], civil safety [8], and industrial production [9]. However, with the development of infrared stealth and camouflage techniques, conventional infrared imaging systems exhibit poor performance in detecting targets in complex environments because of their inability to sense small temperature differences. Infrared polarization imaging [10,11] visually characterizes the polarized information of matter, which offers significant advantages for describing their physical parameters, such as shapes and surface roughness. In target detection applications, we need to accurately identify and locate targets in complex environments. Traditional infrared imaging systems may have limited performance when facing targets with infrared stealth and camouflage technologies. By developing an infrared imaging system with a local polarization channel, we can simultaneously obtain the infrared intensity information and polarization information of the target, thereby improving the accuracy and reliability of target detection. Therefore, as a supplementary tool to traditional intensity imaging, infrared polarization imaging [12,13,14] is capable of identifying objects and their backgrounds in complicated environments, particularly for discriminating artificial targets from natural backgrounds. However, similar to other imaging methods, polarization imaging has certain disadvantages [12,15,16]. For instance, the primary disadvantages of the division of aperture polarimeter [17,18], which is a typical polarization imaging system, are the loss of spatial resolution, larger volume, and weight. Hence, it is necessary for imaging devices to combine the functions of infrared intensity and polarization imaging and employ either of them according to the circumstances. Notably, a locally magnifying imaging system, also called a foveated imager [19], has been catching massive attention. These foveated imagers often work in visible wavelengths and are taken into account in many fields, such as medical instruments [20] and target detection and tracking [21].

In this study, an infrared imaging system with a local polarization channel is proposed. The proposed system satisfies the infrared intensity imaging requirements of a wide field of view (FOV) and acquires magnified polarization images of objects in the field of interest (FOI) [22]. This system employs two infrared imaging channels: a peripheral channel for target searching and tracking and a dynamic local polarization channel for target recognition in the FOI. The two-channel images are on the same plane. An imaging simulation [23,24] was performed using the system, and it demonstrated that the system could acquire local infrared and polarization information of the target while thoroughly monitoring the entire wide FOV. The results presented in this study have significant value for target detection, endoscopy, video surveillance, and agricultural monitoring.

2. Basic Theory

For general passive polarization imaging, it is difficult to acquire high-resolution polarized images over a wide FOV. To address this issue, we developed an infrared imaging system with a local polarization channel to obtain infrared polarization information of the FOI without reducing the entire FOV. As shown in Figure 1, the system consists of three parts: the front imaging lens group; the middle micro-polarization group; and the back relay lens group. The micro-lens group, composed of micro-lens and micro-polarizer, was placed between the imaging lens and relay lens. In Figure 1, the entire target scene is shown in green, and the FOI is shown in red. The peripheral target scene is imaged at the image plane by the imaging lens and relay lens group, which comprise the peripheral imaging channel. The FOI is also imaged at the same image plane but by the imaging lens, micro-polarization group, and relay lens group, which comprise the local polarization channel. The micro-polarization group was placed near the intermediate image plane of the imaging lens group. The difference between the peripheral channel and the local polarization channel is whether the micro-polarization group is passed through by the light.

In this design, the magnification of the local polarization channel, namely, the local magnification, is twice that of the peripheral imaging channel. The micro polarization group was rotated four times, and four images with the polarization characteristics of the FOI, namely, I₀, I₄₅, I₉₀, and I₁₃₅, were obtained. Based on these intensity measurements, three Stokes parameters were obtained [25,26,27], and the relationship was derived using the following equations:

$$I=I_0+I_{90}/2$$

(1)

$$Q=I_0-I_{90}$$

(2)

$$U=I_{45}-I_{135}$$

(3)

where I is the total intensity of the light; Q is the difference between horizontal and vertical polarization, and U is the difference between linear 45° and 135° polarization. The angle of polarization (AOP), that is, the angle of the major axis of the polarization ellipse with respect to the x-axis, can be described as

$$\mathrm{AOP}={\displaystyle \frac{1}{2}}\arctan {\displaystyle \frac{U}{Q}}$$

(4)

Thus, we can obtain the polarization information of the FOI. We can utilize a single optical system to obtain the infrared intensity image of the entire wide FOV and the local polarization image of the FOI in the same image plane.

In addition, when the FOI changes or moves, we can shift the micro-polarization group over the x–y coordinate plane perpendicular to the optical axis to realize dynamic local polarization imaging.

3. System Design

The proposed infrared imaging system with a local polarization channel has a full FOV of 45° with an infrared intensity imaging function and an FOI of 2.6° with a local polarization imaging function. The specifications of the system are listed in

Table 1. Specifications of the infrared imaging system with a local polarization channel.

	EFL/mm	FOV/°	F#	Local Magnification
Peripheral channel	10	45	F/2	/
Local polarization channel	20	2.6	F/4	2×

. Here, to demonstrate the imaging characteristics, three typical FOIs were simulated, namely, Areas 1, 2, and 3. The FOIs of the selected areas were (0°, −1.3°)–(0°, 1.3°), (0°, 8.5°)–(0°, 11.1°), and (0°, 16.4°)–(0°, 19.0°), respectively. These areas covered the FOI from the center to the periphery of the target scene.

Figure 2 shows the infrared imaging system with a local polarization channel designed in this study. The entire effective FOV was 45°, as shown in Figure 2a, whereas the FOI of each area was 2.6°, as shown in Figure 2b–d.

The spot diagrams and modulation transfer function (MTF) plots, which revealed the imaging quality of the peripheral and local polarization channels of the proposed system, are shown in Figure 3. All spot diagrams in the peripheral and local polarization channels are comparable to those of the Airy disk. The MTF, which is the module of the optical transfer function (OTF), characterizes the capability of an imaging system to resolve the observed scene details. The pixel size of the detector used was 25 μm, corresponding to a cut-off spatial frequency of 20 cycles/mm, namely, the Nyquist frequency. The MTF values were greater than 0.25 at 20 cycles/mm. Hence, this imaging system can maintain good image performance in both peripheral and local polarization channels.

4. Imaging Simulation Results

4.1. Local Polarization Imaging

To simultaneously demonstrate the 2× local magnification and local polarization imaging results, both the target scene model shown in Figure 4 and the imaging system model shown in Figure 2 are employed. As shown in Figure 4a, the target scene model was composed of an F-type light source array, wherein several capital letters “F” of the same size in different locations represented the different FOIs in different areas. An extra polarizer or quarter-wave plate was installed next to the F-type light source array to simulate the polarization characteristics of the target.

First, the target scene with linear polarization FOI was simulated. A linear polarizer was placed next to the center “F” to simulate the horizontal linear polarization characteristics of the FOI in the target scene, as shown in Figure 4b; essentially, Area 1 was horizontally polarized. We can obtain four images on the detector with polarization information for the four polarization states of 0°, 45°, 90°, and 135° by rotating the micro-polarizer in the model four times.

As a comparison, the imaging simulation of the infrared imaging system without the microlens group was carried out. The image is shown in Figure 5. The microlens group is a crucial part of the optical system to achieve local magnification and polarization imaging. Without it, the function of the 2× local magnification and local polarization imaging is unachievable. Thus, the light intensity of every “F” in the image plane was equal.

The central “F” was made as Area 1. When the polarization orientation of the micro-polarizer in the imaging system was 0°, which was parallel to the polarization state of the central “F” in the target scene, the light intensity of the central “F” in the image plane was equal to that of the other letters, as shown in Figure 6a. When the polarization orientations of the micro polarizer were 45° and 135°, the light intensity of the central “F” in the image plane was lower than that of the other letters, as shown in Figure 6b,d, respectively. When the polarization orientation of the micro polarizer in the imaging system was 90°, which was perpendicular to the polarization state of the central “F” in the target scene, the light intensity of the central “F” in the image plane was zero, as shown in Figure 6c.

We acquired the polarization images of the target scene by combining Equations (1), (2), and (4), as shown in Figure 7. Figure 7a shows the Stokes vector element I, which was the total intensity of the light. As shown in Figure 7a, the “F” in central FOI takes up 46 pixels in the Y direction, and in contrast, the “F”s in peripheral FOV take up 46 pixels in the Y direction. That is, the magnification of the central “F” is twice that of the other “F”s, which demonstrates the 2× local magnification of the imaging system. Figure 7b shows the Stokes vector element Q, and Figure 7c shows the AOP plot. Those two figures show that only the enlarged central “F” is successfully imaged owing to the polarized property of the central “F” and the unpolarized property of the other “F”s. Thus, for the target with linear polarization characteristics, our system simultaneously captured images in the wide FOV and the local polarization images in the FOI.

Subsequently, the target scene was analyzed with an elliptical polarization FOI. A linear polarizer and quarter-wave plate, which produced elliptically polarized light, were placed next to the central “F” to simulate the elliptical polarization characteristics of the FOI in the target scene, as shown in Figure 8; essentially, Area 1 was elliptically polarized. Four images were obtained on the detector with polarization information for the four polarization states 0°, 45°, 90°, and 135° by rotating the micro-polarizer in the imaging system four times, as shown in Figure 9. Area 1 was elliptically polarized, whereas the other areas were not. Therefore, when rotating the micro-polarizer during imaging, the light intensity of the central “F” in the image plane was lower than those of the other “F”s.

In addition, by combining Equations (1), (2), and (4), we acquired the polarization images of the target scene shown in Figure 10. Figure 10a shows the Stokes vector element I, which was the total intensity of the light. As shown in Figure 10a, the magnification of the central “F” is twice that of the other “F”s, which also demonstrates a 2× local magnification of the system. Figure 10b shows the Stokes vector element Q, and Figure 10c shows the AOP plot. These two figures show that only the enlarged central “F” was successfully imaged owing to the polarized property of the central “F” and the unpolarized property of the other “F”s. Thus, for the target with elliptical polarization characteristics, our system simultaneously captured images in the wide FOV and the local polarization images in the FOI.

4.2. Dynamic Local Imaging

Make the “F” adjacent to the central “F” as Area 2. To demonstrate the scannable imaging capability of the local polarization channel, Area 2 was made horizontally polarized, as shown in Figure 11. Figure 12 shows the images on the detector for the four polarization states of 0°, 45°, 90°, and 135°, and Figure 13 shows the polarization images of the target scene when Area 2 is horizontally polarized. Similarly, for the target scene with linearly polarized Area 2, the system simultaneously captured images in the wide FOV and the local polarization images in the FOI. Thus, the local polarization channel in our system has scannable imaging capability.

5. Conclusions

In this study, we succeeded in developing an infrared imaging system with a local polarization channel. We employed a single optical system with two imaging channels to simultaneously perform infrared imaging with a wide FOV and local polarization imaging with FOI. In this design, the local magnification was twice that of the peripheral imaging channel. Hence, the proposed system can achieve target identification through infrared local polarization imaging without compromising the wide FOV for scene monitoring. This system has superior detection and identification capability under all-time and all-weather conditions.