An infrared array camera point target search system, control method and application

Abstract

The present invention belongs to the field of infrared information acquisition and image processing technology, and discloses an infrared line array camera point target search system, control method and application. The output end of a one-dimensional parallel search and clutter suppression device is connected to a morphological operation device, and the output end of the morphological operation device is connected to a point target detection device; the one-dimensional parallel search and clutter suppression device is provided with an imaging data distributor, a parallel data processor and a parameter corrector; the clutter suppression device is provided with a target expansion operator; the point target detection device is provided with a memory, an operator and a judger. Based on the imaging characteristics of the infrared line array camera, the present invention first performs a one-dimensional parallel rapid search to reduce the influence of the non-uniformity of the infrared camera, mark the potential target point, and then perform a morphological operation on the target point. Finally, according to the preset parameters, the temporal and spatial correlation between frames is used to confirm the target. The use of the present invention can significantly improve the signal-to-noise ratio of point targets in infrared remote sensing images and improve the detection rate.

Description

Technical Field

[0001] The invention belongs to the technical field of infrared information acquisition and image processing, and in particular relates to an infrared linear array camera point target search system, a control method and an application thereof. Background Art

[0002] At present, the core component of modern infrared imaging systems is the focal plane array (FPA), which commonly includes linear array infrared detectors and planar array infrared detectors. Infrared cameras based on linear array infrared detectors are widely used in the field of infrared remote sensing because of their mature detector technology and certain advantages in field of view expansion. The cameras are installed on aircraft or satellites to obtain infrared data for ground remote sensing through push-scan imaging or swing-scan imaging.

[0003] Among them, the detection of moving point targets is a hot topic in current infrared remote sensing technology. Since the distance between the imaging system and the target is far enough, the target is only displayed as a few pixels when it is imaged on the focal plane, lacking recognizable structural information. In addition, due to the strong radiation background of the infrared instrument itself and the complex ground clutter, it is difficult to process infrared images containing moving point targets and extract the target, especially when the target signal is relatively weak. In addition, due to the relative motion between the target and the imaging system, there are phenomena such as the target crossing pixels, the target crossing detector submodules, and the target being distributed in pixel areas with large response non-uniformity, resulting in discontinuity of target intensity and centroid between frames in the real-time signal. For weak signals, it is easier to be submerged in background clutter, which leads to difficulties in infrared remote sensing image processing and single-frame target detection.

[0004] Most of the current processing methods for infrared remote sensing images are separated from their imaging characteristics, and the acquired two-dimensional images are directly filtered in the time domain or space domain. Then, by superimposing multiple frames of images, the signal energy is accumulated on the target motion trajectory to improve the signal-to-noise ratio of point targets. However, this method ignores the influence of the non-uniformity of infrared images, especially for large-field infrared cameras, whose light flux is non-uniform, and the non-uniformity of the system also drifts over time. In addition, this method requires the image to be processed sequentially first, and then the motion trajectory of the target is predicted, and then superimposed. Therefore, it requires strong processing capabilities for sequential images and certain prior knowledge of the motion trajectory of the target. The amount of data calculated and stored is large, which is more suitable for the signal processing of array infrared detectors. For the long linear array infrared detector imaging system based on the splicing of submodules, since the working mode is push-scan or swing-scan, the image frame time is long and the data volume is large, the method of directly predicting the target motion trajectory of the sequential image and then improving the signal-to-noise ratio by accumulating multiple frames is not applicable, and the algorithm needs to be optimized according to the characteristics of the linear array infrared camera.

[0005] Through the above analysis, the problems and defects of the existing technology are as follows: the existing method ignores the structural characteristics of the space camera, fails to utilize the scanning imaging characteristics of the long-line detector imaging system, and the amount of data calculated and stored is large when processing sequential images, requiring a large number of frames, and cannot meet the real-time requirements; and single-frame image processing is easily affected by the non-uniformity of the infrared system, resulting in missed targets.

[0006] The difficulty in solving the above problems and defects is as follows: the infrared imaging system based on long-line detectors needs to be corrected for non-uniformity first, and then the subsequent algorithm processing is performed based on the corrected results. The current algorithm ignores the characteristics of the space camera structure and has high requirements for non-uniformity correction. If the non-uniformity does not meet the requirements (generally required to be within 1%), it will lead to missed detection or high false alarms. The non-uniformity of the current space large-field infrared camera is difficult to maintain within 1% for a long time under various working conditions due to the uneven ambient temperature and light flux. On the other hand, the fluctuation of the infrared background always exists, which aggravates the discontinuous "fault" between the system images caused by the non-uniformity, thereby causing the system detection performance to decline.

[0007] The significance of solving the above problems and defects is that the infrared imaging system based on long-line array detectors can greatly reduce the requirements for system non-uniformity after adopting this method, and the same detection algorithm is not affected by the non-uniformity requirements of the data source and background fluctuations within a certain range. For the same target and the same infrared imaging system, this method improves the detection performance of the system. Summary of the invention

[0008] In order to solve the problems existing in the prior art, the present invention provides an infrared array camera point target search system, a control method and an application.

[0009] The present invention is implemented in this way: an infrared array camera point target search system is provided with:

[0010] One-dimensional parallel search and clutter suppression device;

[0011] The output end of the one-dimensional parallel search and clutter suppression device is connected to a morphological operation device, and the output end of the morphological operation device is connected to a point target detection device.

[0012] Furthermore, the one-dimensional parallel search and clutter suppression device is provided with an imaging data distributor, a parallel data processor and a parameter corrector.

[0013] Furthermore, the clutter suppression device is provided with a target expansion operator.

[0014] Furthermore, the point target detection device is provided with a memory, an operator and a judger.

[0015] Another object of the present invention is to provide a control method for the infrared linear array camera point target search system, the control method comprising: a one-dimensional parallel search and clutter suppression device combines the characteristics of the linear array detector in the scanning direction, firstly performs a one-dimensional parallel rapid search in the scanning direction on a single frame image to mark potential target points; a morphological operation device performs morphological operations on the target points to enhance the target; and a point target detection device confirms the target using inter-frame correlation according to preset motion parameters and correlation coefficients.

[0016] A method for searching point targets using an infrared array camera, comprising the following steps:

[0017] S1, construct M infrared array detector submodules, each detector submodule contains N pixels, and the detectors are spliced ​​and applied to the infrared camera;

[0018] S2, activating an infrared camera containing M infrared array detectors to image the target and obtain scanning line data in real time;

[0019] S3, one-dimensional parallel search and clutter suppression module processes data in real time:

[0020] 1) The imaging data distribution unit presets K parallel processing branches, and each scan line data is divided into K blocks, where the value of K is less than M×N and can be divided by M*N;

[0021] 2) On each parallel processing branch, according to the preset sliding window size 1×L, along the scanning direction, the threshold Thi is calculated once each time the sliding window is slid, and the quantized value Pi of the pixel point at L / 2 is compared with it. If it is greater than or equal to Thi, the original quantized value is retained, and if it is less than Thi, it is set to 0;

[0022] The calculation formula of the threshold value Thi is: Thi = α + β × γ.

[0023] Wherein α is the average value of the quantized value of the corresponding pixel point in the L sliding window, β is the root mean square value of the quantized value of the corresponding pixel point in the L sliding window, and γ is the weighting coefficient.

[0024] 3) After calculating once, the sliding window moves one pixel along the pixel direction and the calculation of 2) is repeated again;

[0025] 4) After completing M×N / K moves and calculations in parallel on K branches, move the 1×L sliding window by 1 pixel along the scanning direction on all branches, and repeat the calculation in 2) until the scanning line ends;

[0026] S4, the result of each sliding window operation, whether in the scanning direction or in the pixel direction, if a target is detected, the target point is expanded by 5×5 through morphological calculation;

[0027] S5. Calculate the target centroid (X0, Y0) in the 9×9 area:

[0028]

[0029]

[0030] Where (Xi, Yi) is the coordinate of the pixel in the i-th column and j-th row in the data, P(Xi, Yi) is the quantized value of the pixel in the i-th column and j-th column in the data, m=8, n=8, which is used for calculation of 9×9 area.

[0031] S6. Record the target center of mass coordinates and the quantized value of the center of mass point to buffer area 1, and write them into memory 1.

[0032] S7. After completing the real-time processing of a frame of data F1, the number of target points is counted, and the parameter correction unit updates the sliding window size of the next frame to L*2, and γ is also updated as appropriate.

[0033] S8, continue the calculation of the new frame F2 in S3, import the stored data of the previous frame F0 into buffer 2, compare and calculate the centroid coordinates with the stored data of the current frame F1, and calculate the centroid distance:

[0034] D i =|(X i ,Y i )-(X' p ,Y' q )|

[0035] S9. When the center of mass distance Di satisfies the following formula, it is judged that the target is hit once.

[0036] D i ≤v·T / (H·r)+η

[0037] Where v is the flight speed of the aircraft, T is the frame time, H is the flight altitude, r is the camera angular resolution, and η is the distortion correction value of the optical-mechanical system.

[0038] S10, iteratively running from S1, the number of buffer areas opened is set to be the same as the number of frames F processed between consecutive frames; when the target is judged to be hit three or more times in a row, the center of mass and quantization value of the target are marked and sent to the post-processing module.

[0039] Another object of the present invention is to provide a method for detecting volcanoes, wherein the method for detecting volcanoes uses the infrared array camera point target search system.

[0040] Another object of the present invention is to provide a method for detecting geothermal heat, wherein the method for detecting geothermal heat uses the infrared array camera point target search system.

[0041] Another object of the present invention is to provide a method for detecting groundwater, wherein the method for detecting groundwater uses the infrared array camera point target search system.

[0042] Another object of the present invention is to provide a method for detecting soil temperature, wherein the method for detecting soil temperature uses the infrared array camera point target search system.

[0043] Another object of the present invention is to provide a method for identifying geological structures, wherein the method for identifying geological structures uses the infrared array camera point target search system.

[0044] Combining all the above technical solutions, the advantages and positive effects of the present invention are as follows: based on the imaging characteristics of the infrared array camera, the present invention first performs a one-dimensional parallel rapid search to reduce the impact of the non-uniformity of the infrared camera, marks the potential target points, then performs morphological operations on the target points, and finally confirms the target based on the preset parameters and the spatial-temporal correlation between frames. The present invention can significantly improve the signal-to-noise ratio of infrared remote sensing image point targets and improve the detection rate.

[0045] The present invention effectively solves the influence of residual non-uniformity of long-line infrared detectors, response differences at the joints of detector modules, target cross-pixel, non-uniformity of light flux of large-field-of-view cameras and other factors on target detection, improves the detection rate, takes into account the requirements of fast search, high detection rate and low false alarm rate of infrared cameras, and solves the problem of fast search of point targets by infrared array cameras under complex backgrounds.

[0046] For the imaging system based on long-line infrared detectors, a simulated point target imaging and detection experiment was set on the ground, and the target signal-to-noise ratio was set to 20. When the non-uniformity correction accuracy was within 1%, the single-frame detection probability of the two-dimensional detection method reached 90%, within 5% to 10%, the detection rate reached 56%, and within 10% to 20%, the detection probability was less than 43%. The two-dimensional detection algorithm used in the experiment is:

[0047] 1) Image decorrelation is performed based on a fixed 2D convolutional template, followed by dynamic threshold segmentation;

[0048] 2) Use multiple frames to correlate and superimpose target signals and identify targets based on trajectory correlation;

[0049] After adopting this method, when the non-uniformity correction accuracy of the infrared system is within 1% to 20%, the single-frame detection probability is greater than 94%, while the residual non-uniformity of the infrared detector is approximately 15% when the non-uniformity correction is not performed. Therefore, ground experimental data show that this method can be effectively applied to situations such as large-field-of-view infrared cameras where it is difficult to improve the non-uniformity correction accuracy and which are greatly affected by infrared background radiation. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0051] Figure 1 It is a structural schematic diagram of an infrared array camera point target search system provided by an embodiment of the present invention;

[0052] In the figure: 1. One-dimensional parallel search and clutter suppression device; 2. Morphological operation device; 3. Target detection device; 4. Imaging data distributor; 5. Parallel data processor; 6. Parameter corrector; 7. Target expansion operator; 8. Memory; 9. Operator; 10. Judger. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0054] In view of the problems existing in the prior art, the present invention provides an infrared array camera point target search system, a control method and an application thereof. The present invention is described in detail below in conjunction with the accompanying drawings.

[0055] like Figure 1 As shown, the infrared array camera point target search system provided by the embodiment of the present invention includes: a one-dimensional parallel search and clutter suppression device 1, a morphological operation device 2, a target detection device 3, an imaging data distributor 4, a parallel data processor 5, a parameter corrector 6, a target expansion operator 7, a memory 8, an operator 9, and a judge 10.

[0056] The output end of the one-dimensional parallel search and clutter suppression device 1 of this embodiment is connected to a morphological operation device 2, and the output end of the morphological operation device 2 is connected to a point target detection device 3; the one-dimensional parallel search and clutter suppression device 1 is provided with an imaging data distributor 4, a parallel data processor 5 and a parameter corrector 6; the clutter suppression device 1 is provided with a target expansion operator 7; the point target detection device 3 is provided with a memory 8, an operator 9 and a judge 10.

[0057] The working principle of the present invention is as follows: the one-dimensional parallel search and clutter suppression device 1 combines the characteristics of the linear array detector in the scanning direction, first performs a one-dimensional parallel fast search in the scanning direction on the single-frame image, and marks the potential target points. The morphological operation device 2 performs morphological operation on the target point to enhance the target. The point target detection device 3 uses the inter-frame correlation to confirm the target according to the preset motion parameters and correlation coefficients.

[0058] Specifically:

[0059] S1, construct 4 infrared linear array detector submodules, each detector submodule contains 512 pixels, and the detectors are spliced ​​and applied to the infrared camera, that is, 2048-element linear array detector;

[0060] S2. Enable the infrared camera to image the target and obtain scanning line data in real time. The line scanning period is 500 microseconds, the frame period is 10 seconds, and the number of scanning lines in a single frame is 20,000 lines.

[0061] S3, one-dimensional parallel search and clutter suppression module processes data in real time:

[0062] 1) 32 parallel processing branches are preset in the FPGA, and each scan line data is divided into 32 blocks, each block has 64 elements;

[0063] 2) On each parallel processing branch, according to the preset sliding window size 1×16, along the scanning direction, the threshold Thi is calculated once each time the sliding window is slid, and the quantized value Pi of the pixel at the center point is compared with it. If it is greater than or equal to Thi, the original quantized value is retained, and if it is less than Thi, it is set to 0;

[0064] The calculation formula of the threshold value Thi is: Thi = α + β × γ.

[0065] Where α is the average value of the quantized value of the corresponding pixel in the L sliding window, β is the root mean square of the quantized value of the corresponding pixel in the L sliding window, and γ is the weighting coefficient, and its initial value is set to 0.1.

[0066] 3) After calculating once, the sliding window moves one pixel along the pixel direction and the calculation of 2) is repeated again;

[0067] 4) After completing 64 moves and calculations in parallel on 32 branches, move the 1×16 sliding window by 1 pixel along the scanning direction on all branches, and repeat the calculation in 2) until 20,000 scan lines are completed;

[0068] S4, the result of each sliding window operation, whether in the scanning direction or in the pixel direction, if a target is detected, the target point is expanded by 5×5 through morphological calculation;

[0069] S5. Calculate the target centroid (X0, Y0) in the 9×9 area:

[0070]

[0071]

[0072] Where (Xi, Yi) is the coordinate of the pixel in the i-th column and j-th row in the data, P(Xi, Yi) is the quantized value of the pixel in the i-th column and j-th column in the data, m=8, n=8, which is used for calculation of 9×9 area.

[0073] S6. Record the target center of mass coordinates and the quantized value of the center of mass point to buffer area 1, and write them into memory 1.

[0074] S7. After completing the real-time processing of a frame of data F1, the number of target points is counted. If the number of points exceeds the threshold, the sliding window size of the next frame is updated to L*2=32, and γ is updated to γ*1.1.

[0075] S8, continue the calculation of the new frame F2 in S3, import the stored data of the previous frame F0 into buffer 2, compare and calculate the centroid coordinates with the stored data of the current frame F1, and calculate the centroid distance:

[0076] D i =|(X i ,Y i )-(X' p ,Y' q )|

[0077] S9. When the center of mass distance Di satisfies the following formula, it is judged that the target is hit once.

[0078] D i ≤v·T / (H·r)+η

[0079] Where v is the target flight speed, T is the frame time, H is the flight altitude, and r is the camera angular resolution. η is the optical-mechanical system distortion correction value. For a 800km orbit low-orbit satellite, the angular resolution is 100urad. For a target moving at a speed of 800km / h, η is 5, and the above value is about 50.

[0080] S10, iteratively running from S1, the number of buffer areas opened is set to be the same as the number of frames F processed between consecutive frames; when the target is judged to be hit three or more times in a row, the center of mass and quantization value of the target are marked and sent to the post-processing module.

[0081] For the imaging system based on long-line infrared detectors, a simulated point target imaging and detection experiment was set on the ground, and the target signal-to-noise ratio was set to 20. When the non-uniformity correction accuracy was within 1%, the single-frame detection probability of the two-dimensional detection method reached 90%, within 5% to 10%, the detection rate reached 56%, and within 10% to 20%, the detection probability was less than 43%. The two-dimensional detection algorithm used in the experiment is:

[0082] 3) Image decorrelation is performed based on a fixed 2D convolutional template, followed by dynamic threshold segmentation;

[0083] 4) Use multiple frames to correlate and superimpose target signals and identify targets based on trajectory correlation;

[0084] After adopting this method, when the non-uniformity correction accuracy of the infrared system is within 1% to 20%, the single-frame detection probability is greater than 94%, while the residual non-uniformity of the infrared detector is approximately 15% when the non-uniformity correction is not performed. Therefore, ground experimental data show that this method can be effectively applied to situations such as large-field-of-view infrared cameras where it is difficult to improve the non-uniformity correction accuracy and which are greatly affected by infrared background radiation.

[0085] In the description of the present invention, unless otherwise specified, "plurality" means two or more than two; the orientations or positional relationships indicated by the terms "upper", "lower", "left", "right", "inner", "outer", "front end", "rear end", "head", "tail", etc. are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0086] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A control method for an infrared array camera point target search system, characterized in that: The infrared array camera point target search device is equipped with: One-dimensional parallel search and clutter suppression device; The output end of the one-dimensional parallel search and clutter suppression device is connected to a morphological operation device, and the output end of the morphological operation device is connected to a point target detection device; The control method of the infrared array camera point target search system includes: The first step is to construct M infrared array detector submodules, each of which contains N pixels. The detectors are spliced ​​and applied to the infrared camera. The second step is to enable an infrared camera containing M infrared array detectors to image the target and obtain scanning line data in real time; The third step is to process the data in real time using the one-dimensional parallel search and clutter suppression module: 1) The imaging data distribution unit presets K parallel processing branches, and each scan line data is divided into K blocks, where the value of K is less than M×N and can be divided by M*N; 2) On each parallel processing branch, according to the preset sliding window size 1×L, along the scanning direction, the threshold Thi is calculated once each time the sliding window is slid, and the quantized value Pi of the pixel point at L / 2 is compared with it. If it is greater than or equal to Thi, the original quantized value is retained, and if it is less than Thi, it is set to 0; The calculation formula of the threshold value Thi is: Thi = α + β × γ; Where α is the average value of the quantized value of the corresponding pixel in the L sliding window, β is the root mean square value of the quantized value of the corresponding pixel in the L sliding window, and γ is the weighting coefficient; 3) After calculating once, the sliding window moves one pixel along the pixel direction and the calculation of 2) is repeated again; 4) After completing M×N / K moves and calculations in parallel on K branches, move the 1×L sliding window by 1 pixel along the scanning direction on all branches, and repeat the calculation in 2) until the scanning line ends; The fourth step is to use the result of each sliding window operation, whether in the scanning direction or in the pixel direction, to expand the target point by 5×5 through morphological calculation if the target is detected; Step 5: Calculate the target centroid (X0, Y0) in the 9×9 area: Where (Xi, Yi) is the coordinate of the pixel in the i-th column and j-th row in the data, P(Xi, Yi) is the quantized value of the pixel in the i-th column and j-th column in the data, m=8, n=8, which is used for calculation of 9×9 area; Step 6: Record the target centroid coordinates and the quantized value of the centroid point into buffer 1 and write into memory 1; Step 7: After completing the real-time processing of a frame of data F1, the number of target points is counted, and the parameter correction unit updates the sliding window size of the next frame to L*2, and γ is also updated as appropriate; Step 8: Continue S3 to calculate the new frame F2, and import the stored data of the previous frame F0 into buffer 2, compare the centroid coordinates with the stored data of the current frame F1, and calculate the centroid distance: D i =|(X i ,Y i )-(X'p,Y'q)|; Step 9: When the center of mass distance Di satisfies the following formula, it is judged as hitting the target once; D i ≤v·T / (H·r)+η; Wherein, v is the flight speed of the aircraft, T is the frame time, H is the flight altitude, r is the camera angular resolution; η is the distortion correction value of the optical-mechanical system; Step 10, iterate from step 1, and set the number of cache areas to be the same as the number of frames F processed between consecutive frames; when the target is hit three or more times in a row, mark the target's center of mass and quantization value and send them to the post-processing module.

2. The control method of the infrared array camera point target search system according to claim 1, characterized in that: The dimensional parallel search and clutter suppression device is provided with an imaging data distributor, a parallel data processor and a parameter corrector.

3. The control method of the infrared array camera point target search system according to claim 1, characterized in that: The clutter suppression device is provided with a target expansion operator.

4. The control method of the infrared array camera point target search system according to claim 1, characterized in that: The point target detection device is provided with a memory, a calculator and a judger.

5. A method for detecting a volcano, characterized in that: The method for detecting volcanoes uses the control method of the infrared array camera point target search system according to any one of claims 1 to 4.

6. A method for detecting geothermal heat, characterized in that: The method for detecting geothermal heat uses the control method of the infrared array camera point target search system according to any one of claims 1 to 4.

7. A method for detecting groundwater, characterized in that: The method for detecting groundwater uses the control method of the infrared array camera point target search system according to any one of claims 1 to 4.

8. A method for detecting soil temperature, characterized in that: The method for detecting soil temperature uses the control method of the infrared array camera point target search system according to any one of claims 1 to 4.

9. A method for identifying geological structures, characterized in that: The method for identifying geological structures uses the control method of the infrared array camera point target search system described in any one of claims 1 to 4.

Images