A bionic multispectral polarization single-pixel imaging device and method based on neural overlapping compound eyes

Abstract

The present invention relates to a bionic multispectral polarization single-pixel imaging device and method based on a neural overlapping compound eye. The device comprises: a substrate; multiple sub-eyes; the multiple sub-eyes are evenly distributed on the substrate, each sub-eye being composed of at least one optical fiber; at least one set of imaging channels, the front end of each set of imaging channels being composed of optical fibers located in the same position in each sub-eye, and the rear end being composed of optical fibers located in the same position in each sub-eye fixed together; at least one filter or polarizer located at the rear end of the imaging channel; and at least one single-pixel detector located behind each filter or polarizer. The imaging device provided by the present invention has a compact structure, small footprint, low production cost, a large imaging field of view, a high imaging dynamic range, and high target recognition efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of bionic compound eyes, single-pixel imaging and target recognition, and in particular to a bionic multi-spectral polarization single-pixel imaging device and method based on neural overlapping compound eyes. Background Art

[0002] Target material identification is of great significance and application value in military reconnaissance, deep space exploration, environmental monitoring, resource assessment, forensic identification, biomedicine, and other fields. Commonly used methods include radar, sonar, infrared imaging, multispectral imaging, polarization imaging, etc. Each of these methods has its own advantages and limitations.

[0003] For example, before using radar for target identification, electromagnetic waves must first be emitted toward the target. This is not only highly susceptible to interference from various external factors, but also incapable of passive target identification. Sonar uses sound waves in water for detection, which can be affected by various conditions in the seawater. While infrared imaging technology itself is not susceptible to interference and can achieve real-time monitoring throughout the day, infrared stealth technology targeting it has also developed rapidly, hindering its further and wider application. The core of spectral recognition technology is that different substances have different spectra, but this method often results in different objects having the same spectrum. Although polarization detection is unaffected by external environmental changes and interference, it lacks the ability to accurately detect targets with low contrast.

[0004] To improve and enhance target recognition, many researchers have proposed a target recognition technique that combines multispectral imaging and polarization imaging, known as spectropolarimetry. This technique can acquire two-dimensional spatial information, one-dimensional spectral information, and polarization state information of a target. Currently available spectropolarimetry imagers are primarily improvements on existing imaging spectrometers and can be categorized into two types: time-sharing imaging and partitioned imaging. Typical examples of time-sharing imaging include spectropolarimetry imagers that use rotating filters and polarizers for imaging, spectropolarimetry imagers based on acousto-optic / liquid crystal tunable filters, and Fourier spectropolarimetry imagers. These techniques acquire multiple types of target information at the expense of temporal resolution. Partitioned imaging, such as the recently proposed computed tomography spectropolarimetry imagers and compressed spectropolarimetry imagers, acquires all spectral and polarization information in a single exposure. However, these techniques employ complex algorithms, require long resolution times, and sacrifice spectral accuracy.

[0005] All of the aforementioned spectral polarimetric imagers are improvements upon traditional single-aperture cameras, and thus suffer from many of the same limitations as conventional cameras. These include a narrow field of view (FOV), which introduces significant image distortion as it increases, a narrow dynamic range, periods of inoperability during deep space exploration, and large image data volumes, requiring high data transmission requirements. Furthermore, all spectral polarimetric imagers, to date, have suffered from low efficiency. Summary of the Invention

[0006] In order to solve the above technical problems, the present invention provides a bionic multi-spectral polarization single-pixel imaging device and method based on neural overlapping compound eyes.

[0007] The technical solution of the present invention is: a bionic multi-spectral polarization single-pixel imaging device based on neural overlapping compound eyes, comprising:

[0008] substrate;

[0009] A plurality of sub-eyes; the plurality of sub-eyes are evenly distributed on the substrate, and each sub-eye is fixed by a plurality of optical fibers;

[0010] Multiple groups of imaging channels, wherein the front end of each group of imaging channels is composed of optical fibers at the same position in each sub-eye, and the back end is composed of optical fibers at the same position in each sub-eye fixed together;

[0011] At least one filter or polarizer, located at the rear end of each group of imaging channels;

[0012] and at least one single-pixel detector located behind each of the filters or polarizers.

[0013] Compared with the prior art, the present invention has the following advantages:

[0014] The present invention discloses a bionic multispectral polarization single-pixel imaging device based on a neural overlapping compound eye. By constructing a single-pixel optical path, single-pixel imaging of a target object is performed. Multi-layer spatial information of the object is obtained by using a compound eye system in conjunction with filters and polarizers. The target can be identified with the help of simple calculations. The device has a compact structure, small footprint, low production cost, large imaging field of view, high imaging dynamic range and high target recognition efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Schematic diagram of the structure of a bionic multi-spectral polarization single-pixel imaging device based on a neural overlapping compound eye in an embodiment of the present invention;

[0016] Figure 2 Schematic diagram of the position of optical fibers in the sub-eye according to an embodiment of the present invention;

[0017] Figure 3 Schematic diagram of the optical path in a bionic multi-spectral polarization single-pixel imaging method based on neural overlapping compound eyes in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The present invention provides a bionic multispectral polarization single-pixel imaging device based on neural overlapping compound eyes, which has a compact structure, small footprint, low production cost, large imaging field of view, high imaging dynamic range and high target recognition efficiency.

[0019] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below through specific implementation and in conjunction with the accompanying drawings.

[0020] Example 1

[0021] like Figure 1 As shown, an embodiment of the present invention provides a bionic multispectral polarization single-pixel imaging device based on a neural overlapping compound eye, comprising:

[0022] Base 101; the base may be a plane, a curved surface, or other shapes. In the embodiment of the present invention, the base adopts a hemispherical curved surface shape;

[0023] Multiple sub-eyes 102; multiple sub-eyes are evenly distributed on the substrate, each sub-eye is fixed by at least one optical fiber; when the sub-eye is composed of multiple optical fibers, each optical fiber in each sub-eye is numbered, such as Figure 2 As shown, it is necessary to ensure that the optical fiber numbers at the same position in each sub-eye are consistent; and the front and back end faces of all optical fibers must be ground and polished;

[0024] At least one group of imaging channels 103, the front end of each group of imaging channels is composed of optical fibers in the same position in each sub-eye, that is, optical fiber No. 1 of each sub-eye is used as the front end of the first group of imaging channels, optical fiber No. 2 of each sub-eye is used as the front end and back end of the second group of imaging channels, and so on; the back end of each group of imaging channels is composed of optical fibers in the same position in each sub-eye fixed together, that is, the back end of optical fiber No. 1 in each sub-eye is fixed as a group, the back end of optical fiber No. 2 is fixed as a group, and so on. When a sub-eye contains n optical fibers (n ≥ 1), n ​​groups of imaging channels can be obtained, and the number of optical fibers contained in each group of imaging channels is the total number of sub-eyes, and the front end of each group of imaging channels is evenly distributed on the substrate along with each sub-eye, and the back end is composed of optical fibers with the same position number in each sub-eye fixed together, and the front and back end surfaces of the optical fibers in each group of imaging channels need to remain flat and consistent;

[0025] At least one filter or polarizer 104 is located at the rear end of each imaging channel; a filter or a polarizer is placed behind each imaging channel; when there are multiple imaging channels, filters or polarizers can be placed at the rear end of all imaging channels as required, or filters or polarizers of different proportions can be placed; Figure 2As shown, in the embodiment of the present invention, each sub-eye contains 7 optical fibers, forming 7 imaging channels. Three polarizers and four filters are placed after the imaging channels, corresponding to the three polarization channels and four spectral channels respectively.

[0026] And at least one single-pixel detector 105 is located behind each filter or polarizer, and is used to collect the light intensity transmitted by each group of imaging channels.

[0027] The bionic multispectral polarization single-pixel imaging device based on neural overlapping compound eyes proposed in the present invention performs single-pixel imaging of the target object by building a single-pixel optical path, and uses the compound eye system in conjunction with filters and polarizers to obtain multi-layer spatial information of the target object. This device has a compact structure, occupies a small space, uses plastic optical fiber as raw material, and is low in cost.

[0028] Example 2

[0029] Based on the device provided in Example 1, a bionic multi-spectral polarization single-pixel imaging method based on neural overlapping compound eyes is realized, and its optical path diagram is as follows: Figure 3 As shown: light source emitter 201, spatial light modulator 202, object to be measured 203, bionic multi-spectral polarization single-pixel imaging device 204 based on neural overlapping compound eye, imaging channel 103, filter and polarizer array 104, single-pixel detector array 105;

[0030] The specific steps are as follows:

[0031] S1: Obtain the mask pattern required for single-pixel imaging and upload it to the spatial light modulator;

[0032] S2: The object to be tested is placed in a light-proof environment. The light emitted by the light source emitter is modulated by the spatial light modulator and reflected onto the object to be tested. The sub-eyes of the bionic multispectral polarization single-pixel imaging device based on the neural overlapping compound eye collect the total light intensity generated by the interaction between the mask pattern and the object to be tested and transmit it through the imaging channel.

[0033] For single-pixel imaging, a light source with a wide bandwidth is required. The bandwidth of the selected light source must encompass the spectra of the selected multiple filters. In this embodiment of the present invention, coal and stone are placed together as the objects to be measured in a light-proof environment. Light emitted by a light source is modulated by a spatial light modulator and reflected onto the object.

[0034] Let P i (x, y) is the mask pattern used for the i-th measurement, where i = 1, 2, 3, ..., M, M is the total number of measurements, and I (x, y) represents the spatial coordinates of the object to be measured. iAfter the total light intensity generated by the interaction of (x, y) and I(x, y) is collected by the sub-eye of the bionic multi-spectral polarization single-pixel imaging device based on the neural overlapping compound eye, since each sub-eye in the embodiment of the present invention is composed of 7 optical fibers, such as Figure 2 As shown, the total light intensity is divided into 7 parts, which are transmitted by each imaging channel respectively;

[0035] S3: Using a filter or polarizer, or a combination of a filter and a polarizer, the total light intensity is modulated and then enters a single-pixel detector to obtain light intensity information;

[0036] In the embodiment of the present invention, an array of four filters and three polarizers is placed at the rear end of the imaging channel. The imaging spectrum of the four filters is 480 (20 nm) nm, 650 (12 nm) nm, 800 (25 nm) nm, and 950 (50 nm) nm, respectively, and the angles of the three polarizers are 0°, 60°, and 120°, respectively. The total light intensity is modulated by the four filters and three polarizers and enters the single-pixel detector, which records the light intensity information, i.e., S i , as shown in formula (1):

[0037] S i =∫∫P i (x,y)I(x,y)dxdy (1)

[0038] S4: Obtain a multispectral image and a polarization image based on the light intensity information; perform band operations and calculate the degree of polarization and polarization angle on the multispectral image and the polarization image respectively;

[0039] Due to S i and P i At this time, (x, y) are all known quantities, so we can calculate and obtain 4 spectra of the object in different bands and 3 polarization images of different polarization states;

[0040] Based on 7 pictures with the same content in 4 different bands and 3 different polarization states, we select the parts of the pictures that we are interested in and perform band calculations and calculations of polarization degree and polarization angle respectively.

[0041] The calculation formulas for the degree of polarization (DOLP) and angle of polarization (AOLP) are as follows:

[0042]

[0043]

[0044] Where I(0°), I(60°), and I(120°) represent the intensities of the linearly polarized components of the light wave at 0°, 60°, and 120°, respectively. I, Q, and U are the first three parameters of the Stokes vector.

[0045] S5: Compare the calculation result with the data in the database to obtain the type of the object to be tested.

[0046] The obtained spectral data of the object to be tested and its polarization degree and polarization angle are compared with the data in the database, and the types of the object to be tested are obtained as coal and stone, thereby completing the target recognition work.

[0047] The bionic multispectral polarization single-pixel imaging method based on neural overlapping compound eyes provided by the present invention can obtain multiple images of the object to be measured in multiple different spectral bands and multiple different polarization angles at one time. These images contain the two-dimensional spatial information of the object to be measured: spectral information and polarization information. After simple processing of the image information, the spatial multi-layer information of the object to be measured can be obtained, and it is compared with the spectra and polarization database of various substances measured in advance, thereby achieving the purpose of material identification of the object to be measured.

[0048] The present invention utilizes single-pixel imaging as its imaging method, which outputs one-dimensional light intensity information. This data volume is far less than that of other spectral and polarimetric imagers, facilitating long-term instrument operation and data transmission. Furthermore, due to the unique characteristics of single-pixel imaging, the two-dimensional spatial information of the object in the resulting multiple images is identical, except for differences in the spectral and polarization information carried. This eliminates the need for subsequent image registration, resulting in improved efficiency.

[0049] The embodiments of the present invention utilize an active single-pixel imaging method, which is theoretically also applicable to passive single-pixel imaging. When using passive single-pixel imaging, since experiments do not require shielding against light, another significant advantage of single-pixel imaging can be realized: a dynamic range far greater than that of conventional cameras. Under varying lighting conditions, imaging over a wide dynamic range can be achieved by simply adjusting the gain of the single-pixel detector. Furthermore, passive single-pixel imaging exhibits outstanding imaging performance in low-light conditions.

[0050] The above embodiments are provided for the purpose of describing the present invention only and are not intended to limit the scope of the present invention. The scope of the present invention is defined by the appended claims. Various equivalent substitutions and modifications made without departing from the spirit and principles of the present invention are intended to be within the scope of the present invention.

Claims

1. A bionic multispectral polarization single-pixel imaging device based on neural overlapping compound eyes, characterized in that: include: substrate; Multiple sub-eyes; A plurality of sub-eyes are evenly distributed on the substrate, and each sub-eye is composed of at least one optical fiber; At least one group of imaging channels, wherein the front end of each group of imaging channels is composed of optical fibers at the same position in each sub-eye, and the rear end is composed of optical fibers at the same position in each sub-eye fixed together; At least one filter or polarizer, located at the rear end of each group of imaging channels; and at least one single-pixel detector located behind each of the filters or polarizers.

2. The bionic multi-spectral polarization single-pixel imaging device based on neural overlapping compound eyes according to claim 1, characterized in that: The base is in a curved shape.

3. A bionic multispectral polarization single-pixel imaging method based on neural overlapping compound eyes, characterized in that: The bionic multispectral polarization single-pixel imaging device based on neural overlapping compound eyes according to any one of claims 1 to 2 comprises the following steps: S1: Obtain the mask pattern required for single-pixel imaging and upload it to the spatial light modulator; S2: The object to be measured is placed in a light-proof environment. The light emitted by the light source emitter is modulated by the spatial light modulator and reflected onto the object to be measured. The sub-eyes of the bionic multi-spectral polarization single-pixel imaging device based on the neural overlapping compound eye collect the total light intensity generated by the interaction between the mask pattern and the object to be measured, and transmit it through the imaging channel. S3: Using a filter or polarizer, or a combination of a filter and a polarizer, the total light intensity is modulated and then enters a single-pixel detector to obtain light intensity information; S4: Obtain a multispectral image and a polarization image according to the light intensity information; perform band operation and calculation of polarization degree and polarization angle on the multispectral image and the polarization image respectively; S5: Compare the calculation result with the data in the database to obtain the type of the object to be measured.

Images