Optical imaging communication method for long-distance large-view-field unmanned aerial vehicle

Abstract

The invention provides a long-distance large-field-of-view unmanned aerial vehicle optical imaging communication method, which comprises the following steps that: a plurality of information sources encode information into a data frame in a specific format, and the encoded data frame is utilized to carry out OOK modulation on optical signals of a plurality of lasers on a transmitting end unmanned aerial vehicle, the plurality of lasers send the modulated optical signals to the orientation of the receiving end unmanned aerial vehicle, and the plurality of lasers are in one-to-one correspondence with the information sources; a plurality of imaging detectors on the receiving end unmanned aerial vehicle perform target detection on a position area of the transmitting end unmanned aerial vehicle in a view field to obtain a detection image, and each imaging detector on the receiving end unmanned aerial vehicle demodulates information transmitted by an optical signal of a corresponding laser on the transmitting end unmanned aerial vehicle according to the image received by the imaging detector, a narrow-band optical filter and a short-focus lens are assembled in front of the imaging detector, the narrow-band optical filter enables the imaging detector to only receive optical signals with the same wavelength as the central wavelength of the narrow-band optical filter, and the short-focus lens enables the imaging detector to obtain a large field of view.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless optical communication, and in particular to a long-distance large-field-of-view UAV optical imaging communication method. Background Art

[0002] Free-space optical communication has unique advantages such as high speed, low power consumption, and strong anti-electromagnetic interference ability. It is one of the important solutions to overcome the congestion of the radio frequency spectrum. However, due to the narrow laser beam at the transmitting end, it is difficult to be detected by the photodetector at the receiving end. Therefore, free-space optical communication generally requires a servo mechanism to control the beam pointing, which is relatively complex and difficult to execute. At the same time, the photodetector cannot separate multiple overlapping optical signals, so it cannot transmit multiple signals.

[0003] Optical imaging communication is an efficient and feasible method with great application prospects. Its characteristic is that it uses an image sensor as the receiver. Therefore, this method has a very high spatial resolution and can receive optical signals within a large spatial range. However, limited by the frame rate of the receiving camera, this method can only be used to transmit low-speed information.

[0004] Free-space optical communication can achieve long-distance point-to-point communication, while optical imaging communication has the potential for large-field-of-view communication. Combining the two with complementary advantages is expected to achieve long-distance large-field-of-view optical communication. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, the present invention provides a long-distance large-field-of-view UAV optical imaging communication method, including the following steps:

[0006] Multiple information sources encode the transmission information into data frames in a specific format, and use the encoded data frames to perform OOK modulation on the optical signals of multiple lasers on the transmitting UAV. The multiple lasers send the modulated optical signals towards the direction of the receiving UAV, where the multiple lasers correspond to the information sources one by one;

[0007] Multiple imaging detectors on the receiving UAV perform target detection on the position area of the transmitting UAV within the field of view to obtain a detection image. Each imaging detector on the receiving UAV demodulates the information transmitted by the optical signal of the corresponding laser on the transmitting UAV according to the received image. A narrowband filter and a short-focus lens are assembled in front of the imaging detector. The narrowband filter enables the imaging detector to only receive optical signals with a wavelength the same as the central wavelength of the narrowband filter, and the short-focus lens enables the imaging detector to obtain a larger field of view.

[0008] Optionally, the multiple information sources encoding the transmission information into data frames in a specific format includes:

[0009] Each information source encodes the transmitted information into a binary sequence with a fixed number of bits as information bits, generates binary check bits with a fixed number of bits based on the information bits, adds the check bits after the information bits, adds 1-bit binary 1 at the very front of the synthesized binary information sequence as the start bit, and finally adds 1-bit binary 1 as the end bit to form a frame structure of start bit, information bits, check bits, and end bit.

[0010] Further, the generating of the binary check bits with a fixed number of bits based on the information bits includes:

[0011] (1) Set 16-bit binary CRC data, initialized to 0XFFFF;

[0012] (2) Perform an exclusive OR calculation on the data of the first byte of the information bits and the 16-bit CRC data, and place the obtained result in the CRC data;

[0013] (3) Shift the CRC data one bit to the right, fill the highest bit with 0, and detect whether the shifted-out data is 0 or 1;

[0014] (4) If the shifted-out data is 0, repeat step (3) again, shift the CRC data one bit to the right and then detect. If the shifted-out data is 1, perform an exclusive OR on the CRC data and the hexadecimal data 0XA001;

[0015] (5) Repeat steps (3) and (4) until all 8 bits of data within the first byte of the information bits are processed;

[0016] (6) Repeat steps (2) to (5) to process the data of the remaining bytes of the information bits;

[0017] (7) Swap the high and low bytes of the finally obtained 16-bit CRC data to obtain the binary check bits.

[0018] Optionally, each imaging detector on the receiving-end unmanned aerial vehicle demodulates the information transmitted by the optical signal of the corresponding laser on the transmitting-end unmanned aerial vehicle according to the received image, including:

[0019] Each imaging detector on the receiving-end unmanned aerial vehicle respectively performs spot target detection on the obtained image to obtain a processed binary image;

[0020] Use a contour recognition algorithm to perform contour recognition on the binary image, calculate the number of contours. If the number of contours is 0, it is considered that there is no spot target in the field of view, and continue to perform target detection until the number of contours is 1, then it is considered that a spot target appears in the field of view;

[0021] When a spot target appears in the field of view, the start position is detected. Set the information count variable and the information content sequence, and set the value of the count variable to 0. Clear the sequence content. For each subsequent frame of the acquired image, perform spot target detection according to the methods of the first two steps. If no spot target is detected, it is considered that the transmitted information bit is 0. Add a 0 to the information content sequence, and increment the count variable by 1. If a spot target is detected, it is considered that the transmitted information bit is 1. Add a 1 to the information content sequence, and increment the count variable by 1. Continue the detection until the value of the count variable reaches the sum of the number of information bits, parity bits, and end bits in the data frame;

[0022] Analyze the received information. If the end bit is 1 and the parity bit data calculated based on the information bit data in the information content sequence is the same as the parity bit data in the information content sequence, the information is transmitted correctly. Receive the information, and restore the information bit data to the transmitted data according to the coding rule. Otherwise, the received information is incorrect and the information is discarded.

[0023] Further, each imaging detector of the receiving UAV performs spot target detection on the acquired image respectively to obtain a processed binary image, including:

[0024] Each imaging detector of the receiving UAV prescribes the minimum pixel length W of the spot diameter when the spot in the received image is valid, and W must be an even number. Let the sliding step be Step, and take a square with side length W as the pixel area size occupied by the target. Let the step Step be Perform spot target detection on the acquired image through the ILCM algorithm. Specifically:

[0025] Let the width of the image acquired by the imaging detector be C and the height be R. Normalize the gray value of each pixel of the image by dividing it by 255.0;

[0026] Judge whether the remainder Crest of C - W divided by Step is 0. If not, expand the width of the image matrix by Step - Crest pixels to the right, and use the pixel values of the rightmost column of the original image to fill the expanded pixel points correspondingly. Then judge whether the remainder Rrest of R - W divided by Step is 0. If not, expand the height of the image matrix by Step - Rrest pixels downward, and use the pixel values of the bottommost row of the original image to fill the expanded pixel points correspondingly to generate an expanded image. Let its width be Cnew and height be Rnew;

[0027] For the expanded image, select the first sub - image block with the upper - left corner as the upper - left vertex and a size of W pixels wide and W pixels high. Then, based on this block, move Step pixels to the right successively, and a total of Sub - image blocks are taken. Then, starting from the first sub - image block in this row, the image block obtained by moving Step pixels downward is used as the next sub - image block. Similarly, it is shifted to the right in turn, and the above operations are repeated until the lower - right vertex of the selected sub - image block coincides with the lower - right vertex of the extended image, obtaining several sub - image blocks. Coordinates of rows and columns are assigned to each sub - image block according to its relative position.

[0028] Let sblk(i, j) be a certain sub - image block, where i and j are the row and column coordinates of the relative position of this sub - image block among all sub - image blocks. Let pix(s, t) be the pixel in the sub - image block, and I(pix(s, t)) be the gray value of the pixel. If the gray - level mean of the sub - image block is m(sblk(i, j)), then Then, the gray - level means of all sub - image blocks are formed into a new matrix M according to the relative positions of the sub - image blocks. The element in the i - th row and j - th column is represented as M(i, j) = m(sblk(i, j)). For the sub - image block sblk(i, j), assume the mean gray value of its pixels is m 0 , and the maximum gray value among all pixels is I max . Taking it as the central sub - image block, find the gray - level means of the 8 adjacent sub - image blocks of sblk(i, j) from the matrix M, denoted as m 1 ~m 8 , and calculate the ILCM of sblk(i, j) as If this sub - image block is an edge block, it is considered that the gray - level means of its missing adjacent sub - image blocks are all 0.0;

[0029] For each element in M, calculate its ILCM respectively, and then arrange the calculation results in the original positions to obtain the saliency map. Multiply the gray value of each pixel in it by 255.0 for denormalization. Then calculate the mean μ SM and standard deviation σ SM , and further calculate the segmentation threshold Th = μ SM + kσ SM , where k is a parameter. The original image is segmented by the segmentation threshold Th.

[0030] After adopting the above technical solutions, the present invention has the following beneficial effects:

[0031] The present invention combines the advantages of free - space optical communication and optical imaging communication, realizes a long - distance and large - field - of - view optical imaging communication method applicable to UAVs, provides a supplementary communication solution when UAV radio frequency communication is limited, and conducts a new exploration of the wireless optical communication method between UAVs. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0033] Figure 1 It is a schematic diagram of the scenario architecture of the communication method proposed by the present invention;

[0034] Figure 2 It is a schematic diagram of the frame format for transmitting information in the present invention;

[0035] Figure 3 It is a schematic diagram of the generation of the extended image of the spot target detection algorithm of the present invention. (a) is the original image obtained by the imaging detector, and (b) is the extended image;

[0036] Figure 4 It is a schematic diagram of the method for selecting sub-image blocks of the ILCM algorithm in the present invention;

[0037] Figure 5 It is a schematic diagram of the sampling of the imaging detector in the present invention;

[0038] Figure 6 It is a schematic diagram of the spot target detection image of the receiving-end unmanned aerial vehicle (UAV) in the present invention. (a) is the image when no spot target appears, and (b) is the image when a spot target appears. Detailed implementation manners

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0040] A single UAV carries multiple lasers with different wavelengths and the corresponding number of imaging detectors. Here, 2 lasers and 2 imaging detectors are taken as examples. The UAV is divided into a transmitting-end UAV and a receiving-end UAV, as Figure 1 shown.

[0041] The two lasers of the transmitting UAV respectively send lasers with wavelengths of 810 nm and 850 nm. After being modulated by two independent information sources respectively, the two lasers are integrated into one signal and sent together, or they can be sent at different times. Due to the same emission aperture but different beam wavelengths, the spot sizes formed by the two signals at the same transmission distance are different. In order to facilitate detection at the receiving end, the optical signals transmitted by the transmitting end should have a large divergence angle. In addition, the information transmission rate of the transmitting end should match the frame rate of the imaging detector at the receiving end. In order to improve the information transmission rate, one frame of image is used to determine one bit of binary data, that is, the transmission rate of the binary code element should be the same as the frame rate of the imaging detector.

[0042] Two large-field-of-view imaging detectors of the receiving UAV are equipped with short-focus lenses. The short-focus lenses enable the imaging detectors to obtain a large field of view. Narrow-band optical filters with central frequencies of 810 nm and 850 nm are respectively installed in front of the lenses, so that optical signals consistent with their central frequencies can be filtered out. The received signal light is projected onto the imaging plane through a lens, and the image type is a single-channel 8-bit grayscale image.

[0043] Based on the above settings, the embodiments of the present disclosure provide a long-distance large-field-of-view UAV optical imaging communication method, including the following steps:

[0044] Two independent information sources encode the transmitted information into data frames in a specific format, specifically as follows:

[0045] Each information source encodes the transmitted information into a binary sequence with a fixed number of bits. For example, it is stipulated that the original information has 10-bit characters. If the transmitted information is less than 10-bit characters, the missing positions are filled with null characters NULL. The character content is encoded in the form of ASCII codes and arranged together. Each 8-bit binary number represents one character, and these data are the information bits;

[0046] Generate check bits with a fixed number of bits according to the information bits, including:

[0047] (1) Set 16-bit binary CRC data, initialized to 0XFFFF;

[0048] (2) Perform exclusive OR calculation on the data of the first byte of the information bits and the 16-bit CRC data, and place the obtained result in the CRC data;

[0049] (3) Shift the CRC data one bit to the right, fill the highest bit with 0, and detect whether the shifted data is 0 or 1;

[0050] (4) If the shifted data is 0, repeat step (3) again, shift the CRC data one bit to the right and then detect. If the shifted data is 1, perform exclusive OR on the CRC data and the hexadecimal data 0XA001;

[0051] (5) Repeat steps (3) and (4) until all 8 bits of data within the first byte of the information bits are completely processed;

[0052] (6) Repeat steps (2) to (5) to process the data of the remaining bytes of the information bits;

[0053] (7) Swap the high and low bytes of the final 16-bit CRC data obtained to get the binary check bits;

[0054] Add the check bits after the information bits, add 1 binary 1 as the start bit at the very front of the synthesized character sequence, and finally add 1 binary 1 as the end bit to form a frame structure of start bit, information bits, check bits, and end bit, as Figure 2 shown.

[0055] Use the encoded data frame to perform OOK modulation on the optical signals of 2 lasers on the transmitting UAV, and the 2 lasers combine the modulated optical signals into 1 path and send them towards the direction of the receiving UAV;

[0056] The 2 imaging detectors on the receiving UAV perform target detection on the position area of the transmitting UAV to obtain a detection image;

[0057] Each imaging detector on the receiving UAV respectively performs spot target detection on the image it obtains to get a processed binary image, including:

[0058] Since the longer the distance between the transmitting and receiving UAVs, the smaller the spot size on the imaging plane of the imaging detector, therefore, each imaging detector on the receiving UAV presets the minimum pixel length W of the spot diameter when the spot in the received image is effective, and W must be an even number to ensure is an integer. Let the sliding step be Step, and use a square with side length W as the pixel area size occupied by the target, and let the step Step be Perform spot target detection on the collected image through the ILCM algorithm, specifically:

[0059] Let the width of the image obtained by the imaging detector be C and the height be R, as Figure 3 (a) shown. Since the image type is single-channel 8-bit, divide the gray value of each pixel of the image by 255.0 for normalization so that the gray value of each pixel is between 0.0 and 1.0, which is convenient for real number operations;

[0060] Determine whether the remainder Crest of C-W divided by Step is 0. If not, expand the width of the image matrix to the right by Step-Crest pixels, and use the pixel values of the rightmost column of the original image to fill the expanded pixel points correspondingly. Then determine whether the remainder Rrest of R-W divided by Step is 0. If not, expand the height of the image matrix downward by Step-Rrest pixels, and use the pixel values of the bottommost row of the original image to fill the expanded pixel points correspondingly, generating an expanded image, as Figure 3 shown in (b). Let its width be Cnew and its height be Rnew;

[0061] For the expanded image, with its upper left corner as the upper left vertex, select the first sub-image block with a size of W pixels wide and W pixels high. After that, based on this block, move Step pixels to the right successively, and a total of sub-image blocks can be selected. Then, using the image block obtained by moving the first sub-image block in this row downward by Step pixels as the next sub-image block, also move it to the right successively and repeat the above operation until the lower right vertex of the selected sub-image block coincides with the lower right vertex of the expanded image, obtaining several sub-image blocks. Assign row and column coordinates to each sub-image block according to its relative position. For example, the row and column coordinates of the first selected sub-image block are both 1, the row and column coordinates of the second one are 1 and 2 respectively, and so on. The range of the row coordinate is 1~ The range of the column coordinate is as Figure 4 shown;

[0062] Let sblk(i,j) be a certain sub-image block, where i and j are the row and column coordinates of the relative position of this sub-image block among all sub-image blocks. Let pix(s,t) be the pixel in the sub-image block, and I(pix(s,t)) be the gray value of the pixel. If the gray mean value of the sub-image block is m(sblk(i,j)), then Then, form a new matrix M with the gray mean values of all sub-image blocks according to their relative positions. Denote the element in the i-th row and j-th column as M(i,j) = m(sblk(i,j)). For the sub-image block sblk(i,j), assume the gray mean value of its pixels is m 0 , and the maximum gray value among all pixels is I max . Taking it as the central sub-image block, find the gray mean values of the 8 adjacent sub-image blocks of sblk(i,j) from matrix M, denoted as m 1 ~m 8 , and calculate the ILCM of sblk(i,j) as If this sub-image block is an edge block, it is considered that the gray mean values of its missing adjacent sub-image blocks are all 0.0;

[0063] For each element in M, calculate its ILCM respectively, and then arrange the calculation results in the original positions to obtain a saliency map. Multiply the grayscale value of each pixel in it by 255.0 for denormalization, and then calculate the mean μ of the saliency map. SM and the standard deviation σ SM , and further calculate the segmentation threshold Th = μ SM + kσ SM , where k is a parameter, and perform binary segmentation on the original image through the segmentation threshold Th;

[0064] Use the contour recognition algorithm to recognize the contours of the binary image and calculate the number of contours. If the number of contours is 0, it is considered that there is no spot target in the field of view, and the target detection is continuously performed until the number of contours is 1, then it is considered that a spot target appears in the field of view;

[0065] When a spot target appears in the field of view, it means that the starting position has been detected. Set the information counting variable and the information content sequence, and set the value of the counting variable to 0, and clear the sequence content. Perform spot target detection on each subsequent acquired image according to the methods of the previous two steps. The corresponding relationship between the sampling frames and the modulation signals is as Figure 5 shown. If no spot target is detected, as Figure 6 (a) shows, it is considered that the transmitted information bit is 0, add a 0 to the information content sequence, and increment the value of the counting variable by 1. If a spot target is detected, as Figure 6 (b) shows, it is considered that the transmitted information bit is 1, add a 1 to the information content sequence, and increment the value of the counting variable by 1, and continue the detection until the value of the counting variable reaches the sum of the number of information bits, parity bits, and end bits in the data frame;

[0066] Analyze the received information. If the end bit is 1 and the parity bit data calculated according to the information bit data in the information content sequence is the same as the parity bit data in the information content sequence, the information transmission is correct, receive the information, and restore the information bit data to the transmitted data according to the coding rule. Otherwise, the received information is incorrect and the information is discarded.

[0067] Although the present invention has been disclosed as above with embodiments, it is not intended to limit the present invention. Any person skilled in the art in the relevant technical field can make some changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to what is defined by the claims.

Claims

1. A long-distance and large-field-of-view unmanned aerial vehicle optical imaging communication method, characterized in that: The following steps are involved: The multiple information sources encode the transmission information into a data frame of a specific format, and use the encoded data frame to perform OOK modulation on the optical signals of the multiple lasers on the transmitting end drone, and the multiple lasers send the modulated optical signals toward the direction of the receiving end drone, wherein the multiple lasers correspond to the information sources one by one; Multiple imaging detectors on the receiving drone perform target detection on the transmitting drone position area within the field of view to obtain a detection image. Each imaging detector on the receiving drone demodulates the information transmitted by the optical signal of the corresponding laser on the transmitting drone according to the image received by each detector. A narrowband filter and a short-focus lens are installed in front of the imaging detector. The narrowband filter allows the imaging detector to only receive optical signals with the same wavelength as the center wavelength of the narrowband filter, and the short-focus lens allows the imaging detector to obtain a larger field of view.

2. The long-distance and large-field-of-view unmanned aerial vehicle optical imaging communication method according to claim 1, characterized in that: The multiple information sources encode the transmission information into a data frame of a specific format, including: Each information source encodes the transmitted information into a binary sequence of a fixed number of bits as information bits, generates a fixed number of binary check bits based on the information bits, adds the check bits to the information bits, adds a binary 1 as the start bit at the front of the synthesized binary information sequence, and finally adds a binary 1 as the end bit, forming a frame structure of start bit, information bit, check bit, and end bit.

3. The long-distance and large-field-of-view unmanned aerial vehicle optical imaging communication method according to claim 2, characterized in that: The step of generating a fixed number of binary check bits according to the information bits includes: (1) Set 16-bit binary CRC data, initialized to 0XFFFF; (2) XOR the data of the first byte of the information bit with the 16-bit CRC data, and place the result in the CRC data; (3) Shift the CRC data right by one bit, fill the highest bit with 0, and check whether the shifted data is 0 or 1; (4) If the shifted-out data is 0, repeat step (3) again, shift the CRC data right one more bit and then check. If the shifted-out data is 1, perform XOR on the CRC data and the hexadecimal data 0XA001. (5) Repeat steps (3) and (4) until all 8 bits of data in the first byte of the information bit are processed; (6) Repeat steps (2) to (5) to process the remaining bytes of information bits; (7) Swap the high and low bytes of the final 16-bit CRC data to obtain a binary check bit.

4. The long-distance and large-field-of-view unmanned aerial vehicle optical imaging communication method according to claim 2, characterized in that: Each imaging detector on the receiving end drone demodulates the information transmitted by the optical signal of the corresponding laser on the transmitting end drone according to the images received by each imaging detector, including: Each imaging detector of the receiving drone performs spot target detection on the image it obtains to obtain a processed binary image; Use the contour recognition algorithm to perform contour recognition on the binary image and calculate the number of contours. If the number of contours is 0, it is considered that there is no spot target in the field of view. Continue to detect targets until the number of contours is 1, then it is considered that a spot target appears in the field of view. When a spot target appears in the field of view, the start bit is detected. The information count variable and information content sequence are set and the count variable value is set to 0. The sequence content is cleared. For each frame of image acquired thereafter, spot target detection is performed according to the methods of the first two steps. If the spot target is not detected, it is considered that the transmitted information bit is 0, and a 0 is added to the information content sequence, and the count variable value is increased by 1. If the spot target is detected, it is considered that the transmitted information bit is 1, and a 1 is added to the information content sequence, and the count variable value is increased by 1. The detection is continued until the count variable value reaches the sum of the number of information bits, check bits, and end bits in the data frame. The received information is analyzed. If the end bit is 1 and the check bit data calculated based on the information bit data in the information content sequence is the same as the check bit data in the information content sequence, the information transmission is correct, the information is received, and the information bit data is restored to the transmission data according to the encoding rules. Otherwise, the received information is incorrect and the information is discarded.

5. The long-distance and large-field-of-view unmanned aerial vehicle optical imaging communication method according to claim 4, characterized in that: Each imaging detector of the receiving end drone performs spot target detection on the image obtained by it to obtain a processed binary image, including: Each imaging detector of the receiving drone predetermines the minimum pixel length W of the effective spot diameter in the received image, and W must be an even number. Let the sliding step length be Step, and the square with a side length of W is the pixel area occupied by the target. Let the step length Step be The ILCM algorithm is used to detect the spot target in the collected image, specifically: Assume that the width of the image acquired by the imaging detector is C and the height is R, and the grayscale value of each pixel of the image is divided by 255.0 for normalization; Determine whether the remainder Crest of CW divided by Step is 0. If not, expand the width of the image matrix to the right by Step-Crest pixels, and use the pixel values ​​of the rightmost column of the original image to fill the expanded pixels. Then determine whether the remainder Rrest of RW divided by Step is 0. If not, expand the height of the image matrix downward by Step-Rrest pixels, and use the pixel values ​​of the bottom row of the original image to fill the expanded pixels. Generate an extended image, set its width to Cnew and height to Rnew; For the extended image, take its upper left corner as the upper left vertex, select the first sub-image block with a width of W pixels and a height of W pixels, and then move right by Step pixels based on this block. A total of sub-image blocks, and then the image block after the first sub-image block in this row is moved down by Step pixels is the next sub-image block, and the same operation is repeated until the lower right vertex of the selected sub-image block coincides with the lower right vertex of the extended image, and several sub-image blocks are obtained. Each sub-image block is assigned row and column coordinates according to its relative position; Let sblk(i,j) be one of the sub-image blocks, i and j be the row and column coordinates of the relative position of this sub-image block in all sub-image blocks, let pix(s,t) be the pixel in the sub-image block, I(pix(s,t)) be the gray value of the pixel, if the gray mean of the sub-image block is m(sblk(i,j)), then Then, the grayscale mean values ​​of all sub-image blocks are combined into a new matrix M according to the relative positions of the sub-image blocks, and the element in the i-th row and j-th column is represented as M(i,j)=m(sblk(i,j)). For the sub-image block sblk(i,j), let the pixel grayscale mean value be m0 and the maximum grayscale value of all pixels be I. max , taking it as the central sub-image block, find the grayscale mean of the 8 adjacent sub-image blocks of sblk(i,j) from the matrix M, denoted as m1~m8, and calculate the ILCM of sblk(i,j) as If the sub-image block is an edge block, it is considered that the grayscale mean values ​​of its missing adjacent sub-image blocks are all 0.0; For each element in M, the ILCM is calculated and the results are arranged in the original position to obtain a saliency map. The gray value of each pixel is multiplied by 255.0 for denormalization, and then the mean μ of the saliency map is calculated. SM and standard deviation σ SM , and then calculate the segmentation threshold Th = μ SM +kσ SM , k is a parameter, and the original image is binarized and segmented using the segmentation threshold Th.

Images