A target detection method and system based on radar and radiometer collaborative perception

Abstract

The application discloses a target detection method and system based on radar and radiometer cooperative sensing, and belongs to the technical field of target detection, which comprises the following steps: S1, acquiring a first source image and a second source image obtained by respectively detecting a same scene by using a radar sensor and a radiometer sensor; S2, after pre-processing the first source image and the second source image, performing target detection to obtain a first target image and a second target image; and S3, after fusing the first target image and the second target image, obtaining the position of a real target. The application considers that the radar sensor has the characteristics of high resolution and distance measurement, and the radiometer sensor has the characteristics of resisting interference from corner reflector and the like, target detection is performed on the imaging results of the radar and the radiometer, and then the radar and the radiometer are fused, the difference of information obtained by different sensors can be fully utilized, complementation between information is realized, the anti-interference capability of detection is improved, and therefore the accuracy of target detection is effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of target detection technology, and more specifically, relates to a target detection method and system based on the coordinated sensing of radar and radiometer. Background Technology

[0002] To cope with complex detection environments, composite detection methods are used to compensate for the shortcomings of single-sensor detection. Composite detectors can leverage the advantages of different sensors, acquire more target scene information, improve the system's anti-interference capability, increase the probability of target capture and data reliability, and more effectively identify target camouflage and deception.

[0003] Existing single-sensor detection methods have low robustness, making it difficult to obtain a true and reliable description of the detection scene and to acquire comprehensive information. Current information fusion methods for detecting complex targets mainly utilize one-dimensional features of the targets, failing to obtain information about their structure and location, thus limiting their effectiveness in target identification. Therefore, the accuracy of existing target detection methods is generally low. Summary of the Invention

[0004] In view of the above-mentioned defects or improvement needs of the existing technology, the present invention provides a target detection method and system based on radar and radiometer collaborative sensing, so as to solve the technical problem of low accuracy of target detection in the existing technology.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a target detection method based on radar and radiometer collaborative sensing, comprising the following steps:

[0006] S1. Acquire the first source image and the second source image obtained by detecting the same scene using a radar sensor and a radiometer sensor, respectively;

[0007] S2. After preprocessing the first source image and the second source image respectively, target detection is performed to obtain the first target image and the second target image;

[0008] S3. After fusing the first target image and the second target image, determine the location of the target area as the result of target detection.

[0009] More preferably, the first target image and the second target image are the same size, both having M rows and N columns; only the detected target area is retained in the first target image and the second target image, and the pixel value of the background area is set to 0.

[0010] More preferably, step S3 above includes:

[0011] S31. Calculate the probability that the pixel in the i-th row and j-th column of the k-th target image belongs to the target region. And the probability that the pixel in the i-th row and j-th column of the k-th target image does not belong to the target region. Where k = 1, 2; Sum represents the number of pixels in the k-th target image whose pixel value is less than the pixel value at the i-th row and j-th column; i = 1, 2, ..., M; j = 1, 2, ..., N; Sum = M × N is the total number of pixels;

[0012] S32. Calculate the conflict influence coefficient between the pixels in the i-th row and j-th column of the first target image and the second target image.

[0013] S33. Calculate the probability that the pixel in the i-th row and j-th column of both the first target image and the second target image belongs to the target region. The confidence function distribution map of the target region is obtained, and thus the location of the target region is determined.

[0014] More preferably, the preprocessing described above includes denoising operations and / or image enhancement operations.

[0015] Further preferably, step S2 includes: preprocessing the first source image and the second source image respectively, and then processing them using the maximum inter-class variance method to obtain the corresponding target mask; performing Hadamard operation on the first source image and the second source image with the corresponding target mask to obtain the first target image and the second target image.

[0016] More preferably, step S1 includes: after acquiring images obtained by using radar sensors and radiometer sensors to detect the same target or batch of targets in the same area, the size of the obtained images is unified to obtain a first source image and a second source image.

[0017] Secondly, the present invention provides a target detection system based on radar and radiometer collaborative sensing, comprising: a radar sensor, a radiometer sensor, and a processing module;

[0018] Radar sensors and radiometer sensors are used to detect the same scene, obtaining a first source image and a second source image;

[0019] The processing module is used to execute the target detection method provided in the first aspect of the present invention.

[0020] Thirdly, the present invention provides a target detection system based on radar and radiometer collaborative sensing, comprising: a memory and a processor, wherein the memory stores a computer program, and the processor executes the target detection method provided in the first aspect of the present invention when executing the computer program.

[0021] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein the computer program, when executed by a processor, controls the device containing the storage medium to perform the target detection method provided in the first aspect of the present invention.

[0022] In summary, the above-described technical solutions conceived in this invention can achieve the following beneficial effects:

[0023] 1. This invention provides a target detection method based on the coordinated sensing of radar and radiometer. Considering that radar sensors have the characteristics of high resolution and range measurement, and radiometer sensors have the characteristics of resistance to interference from corner reflectors, and considering that images contain rich information, by performing target detection on the imaging results of radar sensors and radiometer sensors and then fusing them, the differences in image information obtained by different sensors can be fully utilized, and the information complementarity between the two can be achieved, which improves the anti-interference ability of the detection process and thus effectively improves the accuracy of target detection.

[0024] 2. Furthermore, the target detection method provided by this invention considers that the target image obtained in step S2 only retains pixels of the initially detected target region, while the background region has zero pixels. Moreover, the larger the grayscale value of a pixel, the more likely it is to be a pixel within the target region. Therefore, this invention counts the number of pixels with values ​​smaller than the target region for each pixel and calculates its support for belonging to the target region. The higher the support, the more likely it is to be a pixel within the target region. This yields the probability that each pixel in the first and second target images belongs to or does not belong to the target region, and thus obtains the conflict influence coefficient when the detection results of pixels at the same location in the first and second target images contradict each other. When fusing the first and second target images, the conflict influence coefficient is taken into account, thereby enabling a more accurate determination of the confidence function distribution map of the target region. Based on this, this invention can efficiently fuse information from images detected by different sensors, further improving the accuracy of target detection. Attached Figure Description

[0025] Figure 1 This is a flowchart of the target detection method based on radar and radiometer cooperative sensing provided in Embodiment 1 of the present invention;

[0026] Figure 2 The imaging result of the radar sensor provided in Embodiment 1 of the present invention;

[0027] Figure 3 The imaging result of the radiometer sensor provided in Embodiment 1 of the present invention;

[0028] Figure 4This is the target mask image corresponding to the first source image provided in Embodiment 1 of the present invention;

[0029] Figure 5 This is the target mask image corresponding to the second source image provided in Embodiment 1 of the present invention;

[0030] Figure 6 This is a confidence function distribution diagram of the target region provided in Embodiment 1 of the present invention;

[0031] Figure 7 This is a schematic diagram of the processing flow of the target detection system provided in Embodiment 2 of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0033] Example 1

[0034] A target detection method based on radar and radiometer collaborative sensing, such as Figure 1 As shown, it includes the following steps:

[0035] S1. Acquire the first source image and the second source image obtained by detecting the same scene using a radar sensor and a radiometer sensor, respectively;

[0036] Specifically, after acquiring images of the same target or group of targets in the same area using radar sensors and radiometer sensors respectively, the sizes of the acquired images are unified to obtain the first source image (radar image) and the second source image (synthetic aperture image).

[0037] It should be noted that the imaging results from both the radar sensor and the radiometer sensor are grayscale images, such as... Figure 2 The image shown is the imaging result from the radar sensor; as shown... Figure 3 The image shown is the imaging result of the radiometer sensor.

[0038] S2. After preprocessing the first source image and the second source image respectively, target detection is performed to obtain the first target image and the second target image;

[0039] It should be noted that there are various target detection methods, such as Constant False Alarm Rate (CFAR), Support Vector Machine (SVM), YOLO, and SSD. Preferably, in one optional embodiment, after preprocessing the first source image and the second source image such as denoising and image enhancement, the Otsu's method is used to obtain the corresponding target mask image; the first source image and the second source image are then subjected to Hadamard operation with the corresponding target mask image to obtain the first target image and the second target image.

[0040] In this implementation, Gaussian filtering is used for image denoising; specifically, the two-dimensional Gaussian function is as follows:

[0041]

[0042] Where (x, y) are the pixel coordinates, and σ is the Gaussian filter parameter, which controls the degree of smoothing. A filter with a smaller σ has higher edge localization accuracy but reduces the signal-to-noise ratio of the image; a filter with a larger σ can smooth noise but reduces edge localization accuracy; therefore, the value of σ can be selected according to actual needs.

[0043] The specific denoising process is as follows: establish a coordinate system relationship, calculate the value of the coordinate matrix according to the Gaussian function, normalize the value of the matrix, perform convolution and summation to obtain the gray value of the output image, and assign the result to the output image to obtain the denoised image.

[0044] Furthermore, the specific process of target detection using the Otsu's method in this embodiment is as follows:

[0045] 1) Calculate the normalized histogram of the preprocessed image, using p i Let i = 0, 1, 2, 3, ..., L-1, representing each component of the histogram; L is the number of each component in the histogram.

[0046] 2) Calculate the cumulative sum P1(k), specifically:

[0047]

[0048] i = 0, 1, 2, 3, ..., L-1 represents the components of the histogram.

[0049] 3) Calculate the cumulative mean m(k), k = 0, 1, 2, ..., L-1, specifically:

[0050]

[0051] 4) Calculate the global grayscale value m G Specifically:

[0052]

[0053] 5) Calculate the inter-class variance Specifically:

[0054]

[0055] 6) Obtain the threshold k opt That is, inter-class variance The value of k at the maximum. If the maximum is not unique, then the average of the k values ​​corresponding to each maximum is taken to obtain k. opt .

[0056] 7) Threshold segmentation: Find the optimal threshold k opt Then, the normalized image f(x,y) can be binarized to obtain the corresponding target mask image:

[0057]

[0058] After obtaining the target mask images of the first source image and the second source image through the above steps 1)-7), the first source image and the second source image are subjected to Hadamard operation with the corresponding target mask images to obtain the first target image and the second target image.

[0059] Specifically, the target mask image corresponding to the first source image is as follows: Figure 4 As shown, the target mask image corresponding to the second source image is as follows: Figure 5 As shown. Let the first source image and the second source image be denoted as S respectively. a S b The corresponding target mask images are denoted as T. a T b , target mask image T a T b Perform Hadamard operations on the corresponding source images to obtain the first target image and the second target image F. a F b In this process, only the pixel values ​​of the target area are retained, while the pixel values ​​of the background area are set to 0. The expression is as follows:

[0060] F a =S a ⊙T a

[0061] F b =S b ⊙T b

[0062] The first target image and the second target image are the same size, both having M rows and N columns; only the detected target area is retained in the first target image and the second target image, and the pixel value of the background area is set to 0.

[0063] S3. After fusing the first target image and the second target image, false targets are eliminated, thereby obtaining the position of the real target as the result of target detection.

[0064] Radar sensors are active sensors, offering advantages such as high resolution and range measurement capabilities. However, due to their inherent imaging mechanism, they struggle to distinguish between interference such as corner reflections. Radiometer sensors, on the other hand, are passive sensors. Since they image based on the target's radiation characteristics, they can effectively differentiate between interference and real targets, but their resolution is lower and they cannot perform range measurement. Therefore, the two sensors can complement each other effectively. By performing target detection based on the imaging results from both radar and radiometer sensors and then fusing the results, the accuracy of target detection can be significantly improved.

[0065] It should be noted that algorithms such as additive fusion, multiplicative fusion, multi-scale transformation fusion, principal component analysis fusion, and fuzzy logic fusion can be used for image fusion.

[0066] Preferably, in an optional implementation, image fusion is performed in the following manner:

[0067] S31. For the first target image and the second target image respectively, calculate the probability that each pixel belongs to the target region and does not belong to the target region;

[0068] Specifically, the probability that the pixel in the i-th row and j-th column of the k-th target image belongs to the target region. The probability that the pixel in the i-th row and j-th column of the k-th target image does not belong to the target region. Where k = 1, 2; Sum represents the number of pixels in the k-th target image whose pixel value is less than the pixel value at the i-th row and j-th column; i = 1, 2, ..., M; j = 1, 2, ..., N; Sum = M × N is the total number of pixels;

[0069] It should be noted that for the above target image, since it only retains the pixels of the initially detected target area and the background area has 0 pixels, and the larger the gray value of a pixel, the more likely it is to be a pixel in the target area, this invention counts the number of pixels with a value less than its pixel value for each pixel and calculates its support for the target area. The higher the support, the more likely it is to be a pixel in the target area.

[0070] S32. Calculate the conflict influence coefficient between pixels at the same position in the first target image and the second target image respectively; wherein, the conflict influence coefficient between pixels in the i-th row and j-th column of the first target image and the second target image is:

[0071]

[0072] S33. Calculate the probability that pixels at the same position in the first target image and the second target image belong to the target region, obtain the confidence function distribution map of the target region, and then obtain the location of the target region.

[0073] Specifically, the probability that all pixels in the i-th row and j-th column belong to the target region. The confidence function distribution of the target region is shown in the figure below. Figure 6 As shown.

[0074] The confidence function distribution map of the target region can comprehensively judge the first target image and the second target image and eliminate false targets. Specifically, given a confidence threshold (0.5 in this embodiment), points in the confidence function distribution map that are greater than the confidence threshold are determined as target regions, and points that are less than or equal to the confidence threshold are determined as background regions, thereby obtaining the location of the target region.

[0075] Based on the above method, the present invention efficiently fuses information from images detected by different sensors, increases the amount of information in the fused image, and can better describe the detection scene, thereby greatly improving the accuracy of target detection.

[0076] This invention improves the robustness of collaborative detection using different sensors, enables efficient information fusion of output images from different sensors, avoids the false detection problem caused by using a single sensor in the prior art, has high accuracy, and has high practical application value.

[0077] Example 2

[0078] A target detection system based on radar and radiometer collaborative sensing includes: a radar sensor, a radiometer sensor, and a processing module;

[0079] Radar sensors and radiometer sensors are used to detect the same scene, obtaining a first source image and a second source image;

[0080] The processing module is used to execute the target detection method provided in Embodiment 1 of the present invention.

[0081] Specifically, the processing flow of the aforementioned target detection system is as follows: Figure 7 As shown.

[0082] The relevant technical solutions are the same as in Embodiment 1, and will not be repeated here.

[0083] Example 3

[0084] A target detection system based on radar and radiometer collaborative sensing includes: a memory and a processor, wherein the memory stores a computer program, and the processor executes the target detection method provided in Embodiment 1 of the present invention when executing the computer program.

[0085] The relevant technical solutions are the same as in Embodiment 1, and will not be repeated here.

[0086] Example 4

[0087] A computer-readable storage medium includes a stored computer program, wherein the computer program, when executed by a processor, controls the device where the storage medium is located to execute the target detection method provided in Embodiment 1 of the present invention.

[0088] The relevant technical solutions are the same as in Embodiment 1, and will not be repeated here.

[0089] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A target detection method based on radar and radiometer collaborative perception, characterized in that, The method comprises the following steps: S1, obtaining a first source image and a second source image obtained by respectively detecting a same scene by using a radar sensor and a radiometer sensor; S2, after pre-processing the first source image and the second source image, target detection is performed to obtain a first target image and a second target image; S3, after fusing the first target image and the second target image, the position of a target region is determined as a result of target detection; The first target image and the second target image have the same size, and are both M rows and N columns; only the detected target region is reserved in the first target image and the second target image, and the pixel value of a background region is set to 0; The step S3 comprises: S31. For the first target image and the second target image respectively, calculate the probability that each pixel belongs to the target region and does not belong to the target region; wherein, the probability that the first pixel belongs to the target region and does not belong to the target region. k The first in the target image i Line number j The probability that a pixel in a column belongs to the target region. ;No. k The first in the target image i Line number j The probability that a pixel in a column does not belong to the target region. ; ; For the first k The pixel value in the target image is less than the first i Line number j The number of pixels in the column; ; ; This represents the total number of pixels. S32, calculate the conflict influence coefficient between the pixel points in the first target image and the second target image i row j column ; S33、calculating the probability that the pixel point in the first target image and the second target image belongs to the target region i row j column , obtaining the confidence function distribution of the target region, and further obtaining the position of the target region.

2. The object detection method according to claim 1, wherein The pre-processing comprises a denoising operation and / or an image enhancement operation.

3. The object detection method according to any one of claims 1-2, characterized in that, The step S2 comprises: after pre-processing the first source image and the second source image, a maximum inter-class variance method is used for processing to obtain a corresponding target mask image; the first source image and the second source image are subjected to Hadamard operation with the corresponding target mask image to obtain the first target image and the second target image.

4. The object detection method according to any one of claims 1 to 2, characterized in that, The step S1 comprises: after obtaining images obtained by respectively detecting a same target or a same batch of targets under a same region by using a radar sensor and a radiometer sensor, the sizes of the obtained images are unified to a same dimension to obtain the first source image and the second source image.

5. A target detection system based on radar and radiometer collaborative perception, characterized in that, The method comprises: a radar sensor, a radiometer sensor, and a processing module; The radar sensor and the radiometer sensor are used for detecting a same scene to obtain a first source image and a second source image; The processing module is used for executing the target detection method in any one of claims 1-4.

6. A target detection system based on radar and radiometer collaborative perception, characterized in that, The method comprises: a memory and a processor, the memory stores a computer program, and the processor executes the computer program to execute the target detection method in any one of claims 1-4.

7. A computer readable storage medium characterized by The computer readable storage medium comprises a stored computer program, wherein when the computer program is run by a processor, the device where the storage medium is located is controlled to execute the target detection method in any one of claims 1-4.

Images