Active and passive electromagnetic imaging fusion detection method and system

Abstract

The invention discloses an active and passive electromagnetic imaging fusion detection method and system, and belongs to the technical field of electromagnetic imaging systems and target detection. The objective of the invention is to solve the problems of poor data fusion effect of active imaging and passive imaging and low single-mode target detection precision in the prior art. The method comprises the following steps: acquiring an electromagnetic active imaging picture and an electromagnetic passive imaging picture of a target observation scene by using an imaging radar system; respectively carrying out self-adaptive dual-threshold enhancement on the electromagnetic active imaging picture and the electromagnetic passive imaging picture to obtain an enhanced electromagnetic active imaging picture and an enhanced electromagnetic passive imaging picture; sequentially carrying out superpixel segmentation and region denoising on the enhanced electromagnetic active imaging image and the enhanced electromagnetic passive imaging image to obtain denoised region images, and calculating an image entropy value of each region image; according to the entropy value of each region image, carrying out wavelet fusion on the region images according to a preset proportion to obtain a fused image; and carrying out binary method detection on the fused image to obtain a detection result. The method is suitable for target image detection.

Description

Technical Field

[0001] The present invention relates to the technical fields of electromagnetic imaging systems and target detection technologies. Background Art

[0002] Millimeter-wave (MMW) imaging technology has been widely used in fields such as security screening, monitoring, and medical diagnosis because it can penetrate non-metallic materials (such as clothing) and detect hidden objects. Millimeter-wave imaging technology is mainly divided into two modes: active imaging and passive imaging, each with unique advantages. Active imaging relies on an external illumination source, which can provide a high signal-to-noise ratio (SNR) and system resolution, thereby generating high-resolution images; while passive imaging uses the natural thermal radiation of the target itself and has higher sensitivity to temperature changes and material properties.

[0003] Although passive millimeter-wave imaging has received extensive attention in security screening due to its non-invasive nature and simple structure, its application scenarios have certain limitations. For example, in an indoor environment, passive imaging mainly focuses on target detection against a human background, and its target detection ability in a non-human background is weak. In addition, passive imaging is difficult to effectively distinguish the subtle differences between the target and the surrounding environment, especially when the thermal radiation characteristics of the target and the background are similar. Active imaging uses noise illumination as an external supplementary source, which can enhance the scene detection ability, but it mainly relies on reflection information and ignores the thermal radiation characteristics of the target, which may lead to the loss of important context information. At the same time, active imaging is vulnerable to problems such as specular reflection and frequency resonance, increasing the difficulty of distinguishing the target from the environment.

[0004] To overcome the limitations of a single imaging mode, it is of great significance to fuse active and passive millimeter-wave imaging technologies. Existing research mainly focuses on data-level fusion, usually combining active or passive millimeter-wave images with optical images to improve detection accuracy and interpretability. However, these methods usually use independent active and passive imaging systems, resulting in alignment errors and information inconsistencies between the images, which limits the fusion effect. Although there has been research on developing real-time imaging systems that can simultaneously generate active and passive millimeter-wave images, they mainly focus on analyzing active and passive images separately and fail to fully exploit the potential of their fusion, suffering from poor data fusion effects between active imaging and passive imaging and low single-modal target detection accuracy. Summary of the Invention

[0005] The present invention aims to solve the problems of poor data fusion effects between existing active imaging and passive imaging and low single-modal target detection accuracy, and now provides a method and system for fusing and detecting active and passive electromagnetic imaging.

[0006] The method for fusing and detecting active and passive electromagnetic imaging according to the present invention includes:

[0007] Step S1: Use an imaging radar system to obtain an electromagnetic active imaging map and an electromagnetic passive imaging map of the target observation scene;

[0008] Step S2: Perform adaptive double-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map respectively to obtain the enhanced electromagnetic active imaging map and electromagnetic passive imaging map;

[0009] Step S3: Successively perform superpixel segmentation and regional denoising on the enhanced electromagnetic active imaging map and electromagnetic passive imaging map to obtain the denoised regional images, and calculate the image entropy value of each regional image;

[0010] Step S4: According to the image entropy value of each regional image, perform wavelet fusion on the regional images according to a preset ratio to obtain a fused image;

[0011] Step S5: Perform binary detection on the fused image to obtain a detection result, and complete the active and passive electromagnetic imaging fusion detection.

[0012] Further, in the present invention, in step S1, the method for using an imaging radar system to obtain an electromagnetic active imaging map and an electromagnetic passive imaging map of the target observation scene is as follows:

[0013] According to the detection distance of the radar system, set the integration time, noise wave modulation period, and transmission power of the radar system receiver, control the radar system to transmit the modulated noise wave to the target scene to obtain the target scene detection image, and then collect the modulation signal of the radar system and the target scene detection signal received by the radar system; through a separation algorithm, separate the target detection image to obtain the active image and passive image of the target scene.

[0014] Further, in the present invention, the formula for separating the active image and the passive image through the separation algorithm is:

[0015] T O =||V||

[0016]

[0017] where ||·|| is the L-1 norm operator, T O , T A and T P are the original imaging map, the active imaging map, and the passive imaging map respectively. The matrix V represents the actual voltage output value of the radiometer detector, which is a one-dimensional matrix V, and the matrix M represents the modulation information, which is a one-dimensional matrix M.

[0018] Further, in the present invention, in step S2, the formulas for performing adaptive double-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map respectively to obtain the enhanced images are the same. The formula for performing adaptive double-threshold enhancement on the electromagnetic active imaging map is:

[0019]

[0020] where k represents the k-th gray level, and T UP and T DOWN are respectively the upper threshold and the lower threshold adaptively estimated, is the histogram of the electromagnetic active imaging map after correction, and p(k) is the statistical histogram of the electromagnetic active imaging map.

[0021] Furthermore, in the present invention, in step S3, the process of calculating the image entropy value of each regional image is as follows:

[0022] Step S31: Perform superpixel segmentation on the enhanced active imaging map and the passive imaging map respectively, and divide the images into multiple regional images;

[0023] Step S32: Calculate the saliency value of each regional image based on the gray-scale contrast;

[0024] Step S33: Distinguish the background region and the target region for each regional image according to the saliency value of each regional image, set the brightness temperature value of the background region to 0, and obtain the regional image after removing the background;

[0025] Step S34: Calculate the saliency variance of each regional image after removing the background, and obtain the image entropy value of each region.

[0026] Furthermore, in the present invention, in step S32, the formula for calculating the saliency value of each regional image based on the gray-scale contrast is:

[0027]

[0028] where N(r i ) is the number of pixels in region r i , T i is the average brightness temperature of region r i , and S(r k ) is the saliency value of regional image r k .

[0029] Furthermore, in the present invention, in step S33, the formula for obtaining the regional image after removing the background is:

[0030]

[0031] where α is the saliency threshold, which is dynamically adjusted according to the brightness temperature difference between T A and T P , T(r k ) is the image of region r k after removing the background, and Smax and S min The maximum and minimum values of the saliency values.

[0032] Furthermore, in the present invention, in step S4, the process of obtaining the fused image is as follows:

[0033] Step S41: Perform discrete wavelet transform (DWT) and two-level decomposition on the enhanced active imaging map and the passive imaging map, extract the approximation coefficients and detail coefficients of the active imaging map and the passive imaging map, and capture the horizontal, vertical, and diagonal direction features of the images respectively;

[0034] Step S42: Calculate the fusion weights based on the variance of the regional saliency information; use the fusion weights to perform weighted fusion on the approximation coefficients and detail coefficients of the active imaging map and the passive imaging map;

[0035] Step S43: Perform inverse discrete wavelet transform on the fused approximation coefficients, detail coefficients, and the horizontal, vertical, and diagonal direction features of the image to generate the fused image.

[0036] Furthermore, in the present invention, in step S42, the specific formula for calculating the fusion weights based on the variance of the regional saliency information is:

[0037]

[0038] where ξ a represents the fusion weight of the active imaging map, ξ p represents the fusion weight of the passive imaging map, var(S TP ) and var(S TA ) represent the saliency variances of the passive imaging map and the active imaging map respectively. The larger the saliency variance, the richer the information contained in the image.

[0039] Furthermore, in the present invention, in step S42, the specific formula for performing weighted fusion on the approximation coefficients and detail coefficients of the active imaging map and the passive imaging map using the fusion weights is:

[0040]

[0041] where C f,k represents the k-th order wavelet transform approximation coefficient of the fused image, D f,k represents the k-th order wavelet transform detail coefficient of the fused image, C A,k represents the k-th order wavelet transform approximation coefficient of the active map, D A,k the k-th order wavelet transform detail coefficient of the active map, C P,k represents the k-th order wavelet transform approximation coefficient of the passive map, D P,k represents the k-th order wavelet transform detail coefficient of the passive map.

[0042] Active and passive electromagnetic imaging fusion detection system, which includes: a noise source, an amplifier, a modulator, a transmitting antenna, a focusing lens, a receiving antenna, a radiometer channel, and a data acquisition and processing device;

[0043] The noise source is used to transmit a noise signal to the amplifier;

[0044] The amplifier amplifies the received signal and then sends it to the signal receiving end of the modulator;

[0045] The modulator amplitude-modulates the amplified signal it receives and then sends it to the transmitting antenna, and the transmitting antenna uses the received signal as a detection signal to transmit it to the target scene;

[0046] The focusing lens is used to receive the detection signal reflected by the target scene and focus it on the receiving antenna, and the receiving antenna transmits the received signal to the radiometer channel;

[0047] The radiometer channel is used to generate an observation image of the target scene using the received signal; and send the observation image to the data acquisition and processing device;

[0048] The data acquisition and processing device is used to obtain the electromagnetic active imaging map and the electromagnetic passive imaging map of the target image using the observation image, and perform adaptive dual-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map of the target image, perform superpixel segmentation and regional denoising on the enhanced image, obtain the image entropy value of each regional image, and then perform wavelet fusion on the regional images according to a preset ratio based on the image entropy value of each regional image, and perform binary detection on the fused image to obtain the active and passive electromagnetic imaging fusion detection result.

[0049] Furthermore, in the present invention, the data acquisition and processing device includes a data acquisition module, an image enhancement module, a superpixel segmentation and denoising module, an image fusion module, an image detection module, and an image display module;

[0050] The data acquisition module is used to detect the target scene image and, through a separation algorithm, separate the target detection image to obtain the active image and the passive image of the target scene;

[0051] The image enhancement module is used to perform adaptive dual-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map respectively to obtain the enhanced electromagnetic active imaging map and the electromagnetic passive imaging map;

[0052] The superpixel segmentation and denoising module is used to perform superpixel segmentation and regional denoising on the enhanced electromagnetic active imaging map and the electromagnetic passive imaging map respectively to obtain the denoised regional images, and calculate the image entropy value of each regional image;

[0053] The image fusion module performs wavelet fusion on the regional images according to the image entropy value of each region at a preset ratio to obtain the fused image;

[0054] The image detection module performs binary detection on the fused image to obtain the detection result, completing the active and passive electromagnetic imaging fusion detection.

[0055] Furthermore, in the present invention, the data acquisition and processing device further includes an image display module, and the image display module is used to display the active and passive electromagnetic imaging fusion detection result.

[0056] The present invention combines regional saliency and entropy value adaptive fusion technology, which can enhance the contrast between the target and the background, and at the same time has the characteristics of high resolution and high signal-to-noise ratio of active imaging, and temperature sensitivity and high safety of passive imaging. By fusing the reflection information of active imaging and the radiation information of passive imaging, a high-contrast fused image is generated, thereby improving the target detection ability and environmental adaptability. Brief Description of the Drawings

[0057] Figure 1 is the flowchart of the active and passive electromagnetic imaging fusion detection method described in the present invention;

[0058] Figure 2 is the principle block diagram of the active and passive electromagnetic imaging fusion detection system described in the present invention;

[0059] Figure 3 is the optical picture of a hand-held gun in the implementation case;

[0060] Figure 4 is the original bright temperature image of a hand-held gun in the implementation case;

[0061] Figure 5 is the active image of a hand-held gun in the implementation case;

[0062] Figure 6 is the passive image of a hand-held gun in the implementation case;

[0063] Figure 7 is the passive enhanced and denoised image of a hand-held gun in the implementation case;

[0064] Figure 8 is the active enhanced and denoised image of a hand-held gun in the implementation case;

[0065] Figure 9 is the image fusion result of a hand-held gun in the implementation case;

[0066] Figure 10 is the fusion image detection result of a hand-held gun in the implementation case. Detailed Embodiments

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the 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. It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0068] Specific Embodiment 1: Refer to Figure 1 This specific embodiment will be specifically described. The active and passive electromagnetic imaging fusion detection method described in this embodiment includes:

[0069] Step S1: Use an imaging radar system to obtain an electromagnetic active imaging map and an electromagnetic passive imaging map of the target observation scene;

[0070] Step S2: Perform adaptive double-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map respectively to obtain the enhanced electromagnetic active imaging map and the electromagnetic passive imaging map;

[0071] Step S3: Perform superpixel segmentation and regional denoising on the enhanced electromagnetic active imaging map and the electromagnetic passive imaging map in sequence to obtain the denoised regional images, and calculate the image entropy value of each regional image;

[0072] Step S4: Perform wavelet fusion on the regional images according to a preset ratio based on the image entropy value of each regional image to obtain a fused image;

[0073] Step S5: Perform binary detection on the fused image to obtain a detection result, and complete the active and passive electromagnetic imaging fusion detection.

[0074] Further, in the present invention, in step S1, the method for using an imaging radar system to obtain an electromagnetic active imaging map and an electromagnetic passive imaging map of the target observation scene is:

[0075] According to the detection distance of the radar system, set the integration time, noise wave modulation period, and transmission power of the radar system receiver, control the radar system to transmit the modulated noise wave to the target scene, obtain the target scene detection image, and then collect the modulation signal of the radar system and the target scene detection signal received by the radar system; through a separation algorithm, separate the target detection image to obtain the active image and the passive image of the target scene.

[0076] Further, in the present invention, the formula for separating the active image and the passive image through the separation algorithm is:

[0077] T O = ||V||

[0078]

[0079] Among them, ||·|| is the L-1 norm operator, T O ,T A and T P are the original imaging map, the active imaging map, and the passive imaging map respectively. The matrix V represents the actual voltage output value of the radiometer detector, which is a one-dimensional matrix V. The matrix M represents the modulation information, which is a one-dimensional matrix M.

[0080] Furthermore, in the present invention, in step S2, the electromagnetic active imaging map and the electromagnetic passive imaging map are respectively subjected to adaptive double-threshold enhancement, and the formula for obtaining the enhanced image is the same. The formula for performing adaptive double-threshold enhancement on the electromagnetic active imaging map is:

[0081]

[0082] Among them, k represents the k-th gray level, T UP and T DOWN are the adaptively estimated upper threshold and lower threshold respectively, is the corrected histogram of the electromagnetic active imaging map, and p(k) is the statistical histogram of the electromagnetic active imaging map.

[0083] Furthermore, in the present invention, in step S3, the process of calculating the image entropy value of each regional image is as follows:

[0084] Step S31: Perform superpixel segmentation on the enhanced active imaging map (T A ) and the passive imaging map (T P ) respectively, and divide the image into multiple regional images;

[0085] Step S32: Calculate the saliency value of each regional image based on the gray-scale contrast;

[0086] Step S33: Distinguish the background region and the target region for each regional image according to the saliency value of each regional image, set the brightness temperature value of the background region to 0, and obtain the regional image after removing the background;

[0087] Step S34: Calculate the saliency variance of each regional image after removing the background, and obtain the image entropy value of each region.

[0088] Furthermore, in the present invention, in step S32, the formula for calculating the saliency value of each regional image based on the gray-scale contrast is:

[0089]

[0090] Among them, N(r i ) is the number of pixels in the region r i , Ti is the average brightness temperature of region r i , and S(r k ) is the saliency value of the regional image r k .

[0091] Furthermore, in the present invention, in step S33, the formula for obtaining the regional image after removing the background is as follows:

[0092]

[0093] where α is the saliency threshold, which is dynamically adjusted according to the brightness temperature difference between T A and T P . T(r k ) is the regional image r k after removing the background, and S max and S min are the maximum and minimum saliency values.

[0094] Furthermore, in the present invention, in step S4, the process of obtaining the fused image is as follows:

[0095] Step S41: Perform discrete wavelet transform (DWT) and two-level decomposition on the enhanced active imaging map and passive imaging map, extract the approximation coefficients and detail coefficients of the active imaging map and passive imaging map, and capture the horizontal, vertical, and diagonal direction features of the images respectively;

[0096] Step S42: Calculate the fusion weights based on the variance of the regional saliency information; use the fusion weights to perform weighted fusion on the approximation coefficients and detail coefficients of the active imaging map and passive imaging map;

[0097] Step S43: Perform inverse discrete wavelet transform on the fused approximation coefficients, detail coefficients, and the horizontal, vertical, and diagonal direction features of the image to generate the fused image.

[0098] Furthermore, in the present invention, in step S42, the specific formula for calculating the fusion weights based on the variance of the regional saliency information is as follows:

[0099]

[0100] where ξ a represents the fusion weight of the active imaging map, ξ p represents the fusion weight of the passive imaging map, var(S TP ) and var(S TA ) represent the saliency variances of the passive imaging map and the active imaging map respectively. The larger the saliency variance, the richer the information contained in the image.

[0101] Further, in the present invention, in step S42, the specific formula for weighted fusion of the approximation coefficients and detail coefficients of the active imaging map and the passive imaging map using the fusion weights is as follows:

[0102]

[0103] where C f,k represents the k-th order wavelet transform approximation coefficient of the fused image, D f,k represents the k-th order wavelet transform detail coefficient of the fused image, C A,k represents the k-th order wavelet transform approximation coefficient of the active map, D A,k the k-th order wavelet transform detail coefficient of the active map, C P,k represents the k-th order wavelet transform approximation coefficient of the passive map, D P,k represents the k-th order wavelet transform detail coefficient of the passive map.

[0104] Active and passive electromagnetic imaging fusion detection system, the system includes: noise source 1, amplifier 2, modulator 3, transmitting antenna 4, focusing lens 6, receiving antenna 7, radiometer channel 8 and data acquisition and processing device 9;

[0105] The noise source 1 is used to transmit a noise signal to the amplifier 2;

[0106] The amplifier 2 amplifies the received signal and sends it to the signal receiving end of the modulator 3;

[0107] The modulator 3 amplitude-modulates the amplified signal received and sends it to the transmitting antenna 4, and the transmitting antenna 4 uses the received signal as a detection signal to transmit to the target scene;

[0108] The focusing lens 6 is used to receive the detection signal reflected by the target scene and focus it on the receiving antenna 7, and the receiving antenna 7 transmits the received signal to the radiometer channel 8;

[0109] The radiometer channel 8 is used to generate an observation image of the target scene using the received signal; send the observation image to the data acquisition and processing device 9;

[0110] The data acquisition and processing device 9 is used to obtain the electromagnetic active imaging map and the electromagnetic passive imaging map of the target image using the observation image, and adaptively perform dual-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map of the target image, perform superpixel segmentation and regional denoising on the enhanced image, obtain the image entropy value of each regional image, and then perform wavelet fusion on the regional images according to a preset ratio according to the image entropy value of each regional image, and perform binary detection on the fused image to obtain the active and passive electromagnetic imaging fusion detection result.

[0111] Furthermore, in the present invention, the data acquisition and processing device 9 includes a data acquisition module 91, an image enhancement module 92, a superpixel segmentation and denoising module 93, an image fusion module 94, an image detection module 95, and an image display module 96;

[0112] The data acquisition module 91 is used to detect the target scene image, and through a separation algorithm, separate and obtain the active image and the passive image of the target scene from the target detection image;

[0113] The image enhancement module 92 is used to perform adaptive dual-threshold enhancement on the electromagnetic active imaging map and the electromagnetic passive imaging map respectively, to obtain the enhanced electromagnetic active imaging map and the enhanced electromagnetic passive imaging map;

[0114] The superpixel segmentation and denoising module 93 is used to perform superpixel segmentation and regional denoising on the enhanced electromagnetic active imaging map and the enhanced electromagnetic passive imaging map respectively, to obtain the denoised regional images, and calculate the image entropy value of each regional image;

[0115] The image fusion module 94 performs wavelet fusion on the regional images according to a preset ratio based on the image entropy value of each regional image, to obtain the fused image;

[0116] The image detection module 95 performs binary detection on the fused image to obtain the detection result, and completes the active-passive electromagnetic imaging fusion detection.

[0117] Furthermore, in the present invention, the data acquisition and processing device 9 further includes an image display module 96, and the image display module 96 is used to display the active-passive electromagnetic imaging fusion detection result.

[0118] In this embodiment, a set threshold value is used to perform binary detection on the target.

[0119] The active-passive composite imaging method described in this embodiment is not limited to frequency band limitations, and can be used in fields such as microwave imaging, millimeter-wave imaging, terahertz imaging, etc., and has a wide range of application scenarios.

[0120] Preferably, the modulator can be a single-pole single-throw switch, an amplitude modulator, or other modulation forms, and can also be a method that can produce the same effect, such as time-division transmit on operation and transmit off operation. The noise source can be a thermal radiation noise source, an artificial noise source, etc., and the radiometer channel can be a row of radiometer arrays.

[0121] In this embodiment, as Figures 3 to 9 shown, taking the fusion detection of a handheld concealed item (pistol) in a typical indoor environment as an example, the millimeter-wave active-passive fusion detection method and the active-passive imaging fusion detection system described in the present invention are illustrated. The pistol held by the human body is hidden under the bag, as Figure 3As shown. In this millimeter-wave imaging, the contrast of each item in the active image and the passive image is different. Through the above processing method, the active information and the passive information in the image are separated, and then the active image and the passive image are generated respectively. In the passive image, a highlighted temperature can be clearly observed for the arm part, and in the active image, the handheld item can be detected. The proposed fusion detection algorithm is used to process and fuse the active image and the passive image, and finally the detection result is obtained.

[0122] Figure 4 Shows the original brightness temperature image of the implementation case (handheld gun). Under the noise illumination, the original image can roughly reveal the shape of the pistol, verifying the effective compensation of the noise illumination for the insufficient brightness temperature difference.

[0123] Through the separation algorithm, the active image and the passive image are respectively extracted from the original image. Figure 5 Is the active image of the implementation case (handheld gun), showing the backscattering characteristics of the pistol. However, due to the influence of the specular reflection point, there are some strong point interferences in the image. Figure 6 Is the passive image of the implementation case (handheld gun), mainly reflecting the radiation characteristics of the target, but the presence of the pistol is almost invisible.

[0124] The active image and the passive image are respectively subjected to adaptive double-threshold enhancement, superpixel segmentation and regional denoising processing. Figure 7 Is the passive enhanced and denoised image of the implementation case (handheld gun), where the background noise is effectively suppressed and the target area is clearer. Figure 8 Is the active enhanced and denoised image of the implementation case (handheld gun), where the contour of the pistol is significantly enhanced and the influence of the specular reflection point is weakened.

[0125] Based on the regional saliency information, wavelet fusion is performed on the denoised active image and passive image. Figure 9 Is the image fusion result of the implementation case (handheld gun). The fused image effectively combines the high resolution of the active image and the temperature sensitivity of the passive image, and the contours of both the pistol and the hand are clearly visible.

[0126] The fused image is detected to obtain the final target detection result. Figure 10 Is the detection result of the fused image of the implementation case (handheld gun). The pistol is accurately detected and marked, verifying the effectiveness and practicality of the method.

[0127] So far, all the work of the main-passive electromagnetic imaging fusion detection method and system has been completed. Through the above embodiments, the effect of the main-passive electromagnetic imaging fusion detection method in practical applications is demonstrated. This method can effectively fuse the complementary information of active imaging and passive imaging, significantly improve the accuracy of target detection and the image quality, and is suitable for the detection task of concealed items in complex environments.

[0128] The fusion detection of active and passive electromagnetic imaging in an image is of great significance for target detection in complex scenarios. The method and system for fusion detection of active and passive electromagnetic imaging based on physical characteristics provided by the present invention significantly improve the contrast between the target and the background by fusing the high-resolution information of active imaging and the temperature sensitivity information of passive imaging, breaking through the limitation that some targets are difficult to detect in the traditional single imaging system. This method can effectively combine the backscattering characteristics of active imaging and the radiation characteristics of passive imaging to generate high-quality fused images, providing richer information for target detection. This processing method is simple and efficient, with wide applicability and robustness, and is applicable to various application scenarios such as security screening and concealed object detection.

[0129] Although the present invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed, as long as they do not depart from the spirit and scope of the invention as defined by the appended claims. It should be understood that different dependent claims and the features described herein can be combined in a manner different from that described in the original claims. It should also be understood that the features described in connection with a single embodiment can be used in other described embodiments.

Claims

1. Active and passive electromagnetic imaging fusion detection method, characterized in that: include: Step S1, using an imaging radar system to obtain an electromagnetic active imaging image and an electromagnetic passive imaging image of a target observation scene; Step S2, performing adaptive dual-threshold enhancement on the electromagnetic active imaging image and the electromagnetic passive imaging image respectively to obtain enhanced electromagnetic active imaging image and electromagnetic passive imaging image; Step S3, performing superpixel segmentation and regional denoising on the enhanced electromagnetic active imaging image and the electromagnetic passive imaging image in sequence to obtain a denoised regional image, and calculating an image entropy value of each regional image; Step S4, according to the entropy value of each regional image, performing wavelet fusion on the regional images according to a preset ratio to obtain a fused image; Step S5: perform binary detection on the fused image to obtain detection results, thereby completing active and passive electromagnetic imaging fusion detection.

2. The active and passive electromagnetic imaging fusion detection method according to claim 1, characterized in that: In step S1, the method of obtaining the electromagnetic active imaging image and the electromagnetic passive imaging image of the target observation scene by using the imaging radar system is: According to the detection distance of the radar system, the integration time, noise wave modulation period and transmission power of the radar system receiver are set, the radar system is controlled to transmit the modulated noise wave to the target scene, the target scene detection image is obtained, and then the modulation signal of the radar system and the target scene detection signal received by the radar system are collected; The target detection image is separated into an active image and a passive image of the target scene by a separation algorithm.

3. The active and passive electromagnetic imaging fusion detection method according to claim 2 is characterized in that: Through the separation algorithm, the formula for separating the active image and the passive image is: T O =||V|| Among them, ||·|| is the L-1 norm operator, T O ,T A and T P They are the original imaging image, active imaging image and passive imaging image respectively. The matrix V represents the actual voltage output value of the radiometer detector, which is a one-dimensional matrix V. The matrix M represents the modulation information, which is a one-dimensional matrix M.

4. The active and passive electromagnetic imaging fusion detection method according to claim 3, characterized in that: In step S2, the electromagnetic active imaging image and the electromagnetic passive imaging image are respectively subjected to adaptive dual-threshold enhancement, and the formula for obtaining the enhanced images is the same. The formula for adaptive dual-threshold enhancement of the electromagnetic active imaging image is: Among them, k represents the kth gray level, T UP and T DOWN are the upper and lower thresholds of adaptive estimation, respectively. is the corrected histogram of the electromagnetic active imaging image, and p(k) is the statistical histogram of the electromagnetic active imaging image.

5. The active and passive electromagnetic imaging fusion detection method according to claim 4, characterized in that: In step S3, the process of calculating the image entropy value of each region image is as follows: Step S31, performing superpixel segmentation on the enhanced active imaging image and the passive imaging image respectively, and dividing the image into a plurality of regional images; Step S32, calculating the saliency value of each region image based on the grayscale contrast; Step S33, distinguishing the background area and the target area of ​​each regional image according to the saliency value of each regional image, setting the brightness temperature value of the background area to 0, and obtaining the regional image after removing the background; Step S34: Calculate the saliency variance of each region image after background removal to obtain the image entropy value of each region.

6. The active and passive electromagnetic imaging fusion detection method according to claim 5, characterized in that: In step S32, the formula for calculating the saliency value of each region image based on the grayscale contrast is: Among them, N(r i ) is the region r i The number of pixels, T i For region r i The average brightness temperature, S(r k ) is the region image r k The significance value of .

7. The active and passive electromagnetic imaging fusion detection method according to claim 6, characterized in that: In step S33, the formula for obtaining the regional image after removing the background is: Among them, α is the significance threshold, according to T A and T P The brightness temperature difference is dynamically adjusted, T(r k ) is the area r after removing the background k Image, S max and S min The maximum and minimum significance values.

8. The active and passive electromagnetic imaging fusion detection method according to claim 7, characterized in that: In step S4, the process of obtaining the fused image is as follows: Step S41, performing discrete wavelet transform (DWT) and two-level decomposition on the enhanced active imaging image and passive imaging image, extracting approximate coefficients and detail coefficients of the active imaging image and passive imaging image, and capturing horizontal, vertical and diagonal characteristics of the image respectively; Step S42, calculating fusion weights based on the variance of the regional saliency information; and performing weighted fusion on the approximation coefficients and detail coefficients of the active imaging image and the passive imaging image using the fusion weights; Step S43: Perform inverse discrete wavelet transform on the fused approximation coefficients, detail coefficients, and horizontal, vertical, and diagonal features of the image to generate a fused image.

9. Active and passive electromagnetic imaging fusion detection system, characterized in that: The system comprises: a noise source (1), an amplifier (2), a modulator (3), a transmitting antenna (4), a focusing lens (6), a receiving antenna (7), a radiometer channel (8) and a data acquisition and processing device (9); The noise source (1) is used to transmit a noise signal to the amplifier (2); The amplifier (2) amplifies the received signal and sends it out of the signal receiving end of the modulator (3); The amplified signal received by the modulator (3) is amplitude modulated and then sent to the transmitting antenna (4), and the transmitting antenna (4) transmits the received signal as a detection signal to the target scene; The focusing lens (6) is used to receive the detection signal after being reflected by the target scene and focus it to the receiving antenna (7), and the receiving antenna (7) transmits the received signal to the radiometer channel (8); The radiometer channel (8) is used to generate an observation image of the target scene using the received signal; and send the observation image to the data acquisition and processing device (9); The data acquisition and processing device (9) is used to use the observed image to obtain the electromagnetic active imaging image and the electromagnetic passive imaging image of the target image, and adaptively perform dual-threshold enhancement on the electromagnetic active imaging image and the electromagnetic passive imaging image of the target image, perform super-pixel segmentation and regional denoising on the enhanced image, obtain the image entropy value of each regional image, and then perform wavelet fusion on the regional images according to a preset ratio based on the entropy value of each regional image, perform binary detection on the fused image, and obtain the active and passive electromagnetic imaging fusion detection result.

10. The active and passive electromagnetic imaging fusion detection system according to claim 9, characterized in that: The data acquisition and processing device (9) comprises a data acquisition module (91), an image enhancement module (92), a super-pixel segmentation and denoising module (93), an image fusion module (94), an image detection module (95) and an image display module (96); The data acquisition module (91) is used to detect the target scene image, and separate the target detection image into an active image and a passive image of the target scene through a separation algorithm; The image enhancement module (92) is used to perform adaptive dual-threshold enhancement on the electromagnetic active imaging image and the electromagnetic passive imaging image respectively, to obtain an enhanced electromagnetic active imaging image and an enhanced electromagnetic passive imaging image; The super pixel segmentation and denoising module (93) is used to perform super pixel segmentation and regional denoising on the enhanced electromagnetic active imaging image and the electromagnetic passive imaging image respectively, to obtain a denoised regional image, and to calculate an image entropy value of each regional image; The image fusion module (94) performs wavelet fusion on the regional images according to the entropy value of each regional image at a preset ratio to obtain a fused image; The image detection module (95) performs binary detection on the fused image to obtain a detection result, thereby completing active and passive electromagnetic imaging fusion detection.

Images