A method and system for detecting a target carrying radioactive material

Abstract

The application discloses a kind of target detection method and system carrying radioactive substance.The method of the present application comprises: collecting the visible light image and radioactive image of the target to be measured;Based on the predetermined characteristics of each target to be measured are extracted based on the visible light image;Determine the first trajectory of the target to be measured based on multiple frames of visible light image within a predetermined time, and then assign a first weight w1 to each target to be measured based on the confidence sum value;Determine the second trajectory of radioactive material based on the radioactive image, and determine the second weight w2 of each target to be measured according to the matching similarity of the two trajectories;Determine the target to be measured with radioactive material based on the combination of the two weights.The method of the present application can detect suspicious target personnel carrying radioactive material at a relatively low activity, improve the correctness of radioactive material positioning screening in public places such as security check and customs, significantly reduce the false detection rate, and does not need to increase new hardware, stable and reliable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of nuclear radiation detection, in particular to a radioactive material carrying target detection method and system. BACKGROUND

[0002] Radioactive material detection has its unique application in some public places, such as transportation hubs and large conferences. Gamma cameras are usually used to detect pedestrians or objects carrying radioactive materials, thereby excluding the risk of special nuclear materials and radioactive substances carried by passengers in their luggage or hand luggage, and improving the safety of public places with frequent personnel exchanges.

[0003] In the prior art, the combination of gamma cameras and visible light cameras enables more and more information to be obtained during radioactive material detection. When the activity of radioactive material is high, it can usually be tracked more accurately by a radioactive material positioning system, but for low-activity radioactive materials or long-distance imaging, it is often difficult to accurately locate and effectively track low-count radioactive materials in a timely manner due to the physical limit of the signal-to-noise ratio of the detector. SUMMARY

[0004] In order to solve the problem of tracking and locating radioactive sources in a very low count situation (such as a dose equivalent condition higher than 2 times the background), the present application provides an improved radioactive material carrying target detection method. The method of the present application can still accurately identify targets carrying radioactive sources at low counts by relying on existing conventional radiation detectors, quickly locate and track the path of radioactive sources, and ensure the safety of real-time monitoring areas. Of course, in order to better implement the method of the present application, the present application also provides an improved radioactive material carrying target detection system. The method and system of the present application will be described in detail below.

[0005] Specifically, in one aspect, the present application provides a radioactive material carrying target detection method, comprising:

[0006] Step 1), acquiring a visible light image and a radioactive image of the target to be detected;

[0007] Step 2), extracting predetermined features of each target to be detected based on the visible light image, and determining a first trajectory of the predetermined features;

[0008] Step 3), determining a confidence sum value of the radioactivity corresponding to the predetermined features of each target to be detected, and assigning a first weight w1 to it;

[0009] Step 4), determining a second trajectory of the radioactive material based on the radioactive image, and determining a second weight w2 of each target to be detected according to the matching degree of the first trajectory and the second trajectory of each target to be detected;

[0010] Step 5), determining the target carrying radioactive material to be detected according to the first weight w1 and the second weight w2.

[0011] The first weight w1 is positively correlated with the confidence sum value, for example, the confidence sum value itself can be taken as the first weight w1 value or the confidence sum values of each target to be detected are sorted, and the first weight w1 value is assigned to each target according to the sorting, and the higher the sorting, the greater the weight.

[0012] The second weight w2 is positively correlated with the matching degree, for example, the matching degree itself can be taken as the second weight w2 value or the matching degrees of the first trajectory and the second trajectory of each target to be detected are sorted, and the second weight w2 value is assigned to each target according to the sorting, and the higher the sorting, the greater the weight.

[0013] In a preferred implementation, the method further comprises obtaining a fusion image based on the visible light image and the radioactive image, labeling the target carrying radioactive material to be detected in the fusion image based on the determined target, and the method further comprises acquiring the radioactive count rate of the target region, issuing an alarm signal based on the radioactive count rate threshold, and performing the detection method in steps 1)-5) in response to the alarm signal.

[0014] In another preferred implementation, step (3) comprises mapping the predetermined feature of each frame of visible light image to the radioactive image of the corresponding acquisition time along the first trajectory, determining the confidence sum value of the predetermined feature about radioactivity for each target to be detected, and assigning the first weight w1 to each target based on the confidence sum value.

[0015] In another preferred implementation, the confidence sum value of the predetermined feature about radioactivity is determined by the following formula:

[0016]

[0017] where p is the number of gamma photons detected by each detector pixel j of the radioactive image acquisition device p j is the vector, j is the index of the detector pixel of the radioactive image acquisition device, i is the pixel index after the gamma photon incident angle plane is discretized, f is the vector distribution of the number of gamma photons incident to the front surface of the radioactive image acquisition device, f i is the number of gamma photons incident to the front surface of the radioactive image acquisition device (gamma camera collimator) from the discretized target angle plane, c ij is the system transmission matrix quantitatively describing the response of the radioactive image acquisition device to incident gamma photons.

[0018] In another aspect, the present application provides a radioactive material carrying target detection system, comprising: a visible light image acquisition device, a radioactive image acquisition device, a reconstruction module, a tracking module, a positioning module, and a discrimination module,

[0019] The visible light image acquisition device is configured to acquire visible light images of targets to be detected.

[0020] The radioactive image acquisition device is configured to acquire radioactive count rates of the targets to be detected.

[0021] The reconstruction module is configured to reconstruct radioactive images based on the radioactive count rates.

[0022] The tracking module is configured to extract predetermined features of each target to be detected based on the visible light images, and obtain a first trajectory according to the predetermined features.

[0023] The positioning module is configured to determine a confidence sum value of the predetermined features of each target to be detected with respect to radioactivity, assign a first weight w1 to each target to be detected based on the confidence sum value, and the discrimination module is configured to determine a second trajectory of radioactive material based on a plurality of the radioactive images output by the reconstruction module within a predetermined time, match curves of the first trajectory and the second trajectory of each target to be detected, determine a second weight w2 of each target to be detected based on a matching similarity, and combine the first weight w1 and the second weight w2 corresponding to each target to be detected to obtain a confidence weight w of each target to be detected with respect to radioactivity, and determine the target to be detected with radioactive material based on a weight ranking.

[0024] In a preferred implementation, the reconstruction module is further configured to reconstruct a fusion image based on the visible light images and the radioactive count rates, and the target detection system further comprises a display module configured to mark the target in the fusion image based on a determined position of the target to be detected with radioactive material.

[0025] In another preferred implementation, the positioning module is configured to map the predetermined features of each frame of visible light images to the radioactive images at corresponding times along the first trajectory, determine a confidence sum value of the predetermined features of each target to be detected with respect to radioactivity, and assign a first weight w1 to each target to be detected based on the confidence sum value, preferably, the predetermined features include, but are not limited to, human faces, body parts, or carried objects.

[0026] In another preferred implementation, the radioactive image acquisition device is a gamma camera, which determines a real-time framing frequency according to count rate information of received gamma photons, and reorganizes gamma photon data.

[0027] In another preferred implementation, the positioning module determines the confidence sum value of the predetermined feature with respect to radioactivity using the following formula:

[0028]

[0029] wherein p is the number of gamma photons detected by each detector pixel j of the radioactivity image acquisition device j is a vector, j is the index of the detector pixel of the radioactivity image acquisition device, i is the index of the pixel after the gamma photon incident angle plane is discretized, f is the vector distribution of the number of gamma photons incident to the front surface of the collimator of the radioactivity image acquisition device, f i is the number of gamma photons incident to the front surface of the radioactivity image acquisition device (gamma camera collimator) from the discretized target angle plane, c ij is a system transmission matrix quantitatively describing the response of the radioactivity image acquisition device to incident gamma photons.

[0030] In another preferred implementation, for the radioactivity image acquisition device and the visible light image acquisition device that are installed with a downward angle tilt, the fusion image of the radioactivity image and the visible light image and the depth correspondence relationship in the three-dimensional space are obtained according to the spatial positions at which the radioactivity image acquisition device and the visible light image acquisition device are installed and the fusion registration relationship therebetween, and when the feature points of different targets delivered by the tracking module are obtained, the depth direction of the face is corrected according to the area of the face, and the correction formula is as follows:

[0031] Y' = S / Sd * Y index

[0032] wherein S is the pixel area of the visible light image occupied by the extracted face, Sd is the standard face area corresponding to the Y direction pixel position at which the human eye is located, Y index is the Y direction pixel position of the extracted feature point, and Y' is the Y direction pixel position of the corrected feature point.

[0033] Technical effects

[0034] By using the dual-mode images of the existing conventional radiation detector and the visible light camera, the method and system of the present application can detect and image and fusion display radioactive substances under the condition of low count close to the background, thereby significantly improving the accuracy of system detection and screening. No additional hardware cost is needed, and the input mode is not limited, which is easy to expand and integrate. It is also easier to combine with night vision images, infrared images, etc. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 is a schematic block diagram of the detection system of the present application;

[0036] Figure 2 The correlation between the radioactive material moving track (obtained after reconstruction of multiple frame projection images) and the moving track of the detected feature target (the detected target is set as a human face) is shown;

[0037] Figures 3-5 A radioactive source detection process diagram in a simulation test environment is shown;

[0038] Figure 6 The actual moving track of the radioactive material and the radioactive source track after correction of the depth value according to the human face area are shown. DETAILED DESCRIPTION

[0039] The embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings. The embodiments described below by referring to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be understood as a limitation of the present application.

[0040] As Figure 1 shown, the target detection system of the present embodiment includes a data acquisition module (including a visible light image acquisition device and a radioactive image acquisition device), a reconstruction module, a tracking module, a positioning module and a discrimination module. The shooting areas of the visible light image acquisition device and the radioactive image acquisition device overlap, and the overlapping part is used as the target area for detection. For example, the system can be placed in an airport, a train station, a customs office and the like, and used for detecting the passing personnel.

[0041] (1) Image acquisition

[0042] The visible light image acquisition device is used to acquire the visible light images of each target to be detected in the target area, and the radioactive image acquisition device is used to acquire the radioactive images or radioactive particle, gamma photon count of the target. It should be noted that the system of the present application can be used in association with an existing radioactive detector, at this time, the visible light image acquisition device and the radioactive image acquisition device can be a visible light camera and a gamma camera provided in the radioactive detector. Or the visible light image acquisition device and the radioactive image acquisition device are only used to receive the corresponding visible light images and radioactive images. In the present embodiment, the radioactive acquisition device uses a gamma camera.

[0043] (2) Image reconstruction

[0044] The reconstruction module is used for image reconstruction based on the radioactive images.

[0045] Reconstruction principle: When the radioactive source is located in the far field, the gamma photon beams emitted from the same position point source to the detector area can be approximately considered as parallel beams from a certain angle direction of the two-dimensional angle plane (α, β). The gamma photon beams pass through the opening area of the gamma camera front collimator to the detector, interact with the detector, and after a series of conversions from gamma photons to visible light and then to current signals, the energy spectrum and projection data p of the gamma photons are obtained by the acquisition computer.

[0046] The projection data p is the number of gamma photons p j detected by each detector pixel j of the gamma camera, J is the number of detector pixels of the gamma camera, j = 1, …, J; the target angle plane (α, β) is discretized into I angle plane pixels, f i is the number of gamma photons incident from the discretized (α i , β i ) direction to the front surface of the gamma camera collimator, where i = 1, …, I, and the vector f i composed of f i,j represents the distribution f(α, β) of the flux (or particle fluence) of the incident gamma photon parallel beam in the target angle plane (α, β).

[0047] Using the maximum likelihood estimation statistical method, a quantitative estimate of f can be obtained i.e. a two-dimensional gamma radiation image with quantitative information, as follows

[0048]

[0049] In the formula, M is the system transmission matrix quantitatively describing the response of the gamma camera to incident gamma photons, and the element M i,j of the jth row and ith column of M represents the probability that the gamma photons incident from the (α i , β i ) direction of the target angle plane are detected at the detector pixel j of the gamma camera.

[0050] The gamma camera can obtain the energy, position and time information of the photons incident to the detector, and the photon information in a unit of time can obtain the corresponding gamma photon energy spectrum distribution and projection distribution. According to the system transmission matrix and the reconstruction formula, the distribution of the incident gamma photons can be obtained, and the radioactive distribution map can be obtained.

[0051] According to the gamma photon counting rate information obtained by the gamma camera in real time, the reasonable frame number of the projection image per second is reorganized:

[0052] 1) Counting rate Countrate≤50, 1 frame per second projection;

[0053] 2) Counting rate Countrate≤100 & Countrate>50, 2 frames per second;

[0054] 3) Counting rate Countrate≤150 & Countrate>100, 3 frames per second;

[0055] 4) Counting rate Countrate≤250 & Countrate>150, 4 frames per second;

[0056] 5) Counting rate Countrate>250, 5 frames per second;

[0057] The reconstructed projection images are respectively reconstructed, and the reconstructed images are superimposed to obtain a moving track of the radioactive source on the gamma image; (since the actual radioactive source storage position is complex, the moving track of the different targets extracted by the visible light image does not necessarily conform to the actual radioactive source moving track), since the field of view of the visible light camera is usually larger than that of the gamma camera, there may be a case that the target exceeds the field of view.

[0058] (3) Target tracking

[0059] The tracking module is configured to extract predetermined features of each to-be-measured target based on the visible light images, and determine a first track of a specific point in the predetermined features based on a plurality of frames of visible light images within a predetermined time. The to-be-measured target is usually a person or luggage that may carry a radioactive source. Figures 2-6 The present application is illustrated by taking a person as a to-be-measured target. The features of the person can be a face, a head, a torso, etc. In this embodiment, the face of the person is used as the feature for track recognition. According to the distribution of the feature points at different times, the moving track of the person and the article is obtained. It is generally believed that radioactive substances are carried by the human body or articles.

[0060] Based on a period of time, such as 30s, the center or the barycenter position of the face feature of each person in the visible light image is taken as the track point position of the person, and the visible light images within 30s will form a continuous or discontinuous face center point position track, which is referred to as a first track herein.

[0061] (4) Target positioning

[0062] The positioning module is configured to map each point in the face feature or the entire region of the body of each to-be-measured target in each frame of visible light image to the radioactive image at the corresponding time along the first track, determine a confidence sum value of the face feature or the entire body of each to-be-measured target with respect to radioactivity, and assign a first weight w1 to each to-be-measured target based on the confidence sum value, wherein the confidence sum value is positively correlated with the first weight w1.

[0063] For example, assuming that the body of a person occupies a pixel area of 56*128 in a visible light image. Since the visible light acquisition device and the radioactivity acquisition device are registered with each other, according to the fusion registration relationship from the visible light image to the gamma radiation image, each pixel in the pixel area can be mapped to a corresponding coordinate in the radioactivity distribution map respectively; according to the confidence calculation formula, for each pixel of each target, the confidence values of all feature points in the body of the target with respect to radioactivity are calculated respectively, and the confidence values of all point coordinates are summed to obtain the total confidence value of the person with respect to radioactivity.

[0064] For example, the calculation of the confidence value can be based on the following manner:

[0065] The real-time gamma image moving trajectory is taken as an input image, the position value corresponding to the trajectory is set to 1, and the rest is set to 0. An N*N normalized Gaussian filter window function is set, and is convolved with the gamma image with the trajectory to calculate the corresponding confidence value result.

[0066] In an implementation manner, the following formula is used to determine the total confidence value of a predetermined feature, such as a face, with respect to radioactivity:

[0067]

[0068] wherein p is the number of gamma photons detected by each detector pixel j of the radioactivity image acquisition device p j , j is the index of the detector pixel of the radioactivity image acquisition device, i is the pixel index after the gamma photon incidence angle plane is discretized, f represents the vector distribution of the number of gamma photons incident to the front surface of the radioactivity image acquisition device, c ij is a system transmission matrix for quantitatively describing the response of the radioactivity image acquisition device to the incident gamma photons.

[0069] (5) Target discrimination

[0070] The discrimination module is configured to determine a second trajectory of the radioactive substance based on a plurality of radioactivity images in a predetermined time, or if the second trajectory has been generated in a reconstruction module of an existing radioactivity detection device, the discrimination module directly reads the second trajectory, matches the curves of the first trajectory of each to-be-detected person and the second trajectory, determines a second weight w2 of each to-be-detected target based on a matching similarity, the higher the similarity, the higher the weight, and combines the first weight w1 and the second weight w2 corresponding to each to-be-detected target to obtain a confidence weight w of each to-be-detected target with respect to radioactivity, and determines the to-be-detected target with the radioactive substance based on a weight ranking.

[0071] The similarity matching can be performed by using existing curve similarity calculation methods, such as peak error, absolute error, curve distance similarity, etc. The higher the determined similarity is, the greater the second weight w2 is, and the two are positively correlated.

[0072] The main purpose of the process is to determine the target with the highest possibility of carrying radioactive substances, so as to be displayed in the fusion image subsequently.

[0073] Preferably, the detection system of the present application further comprises a display module for marking the target based on the position of the determined person to be tested with radioactive substances in the visible light image or the fusion image.

[0074] In an implementation manner, the process of fusion display is:

[0075] a) In the statistical time of real-time framing, the confidence sum results corresponding to different targets are calculated according to the confidence set of different targets obtained based on the mapping process from the visible light image to the radioactive image;

[0076] 2) According to the confidence sum results of different targets, the different targets are marked with their corresponding weight value coefficients w1 according to the sorting results;

[0077] 3) Within the time range of this transmission before the end of the alarm, the image reconstructed by the reconstruction module can be optionally counted in real time to obtain the moving track of the radioactive substance;

[0078] 4) According to the moving track of different targets transmitted by the tracking module 30, the center of gravity of the face is taken as the first track, and the curve of the moving track of the radioactive substance (the second track) is matched, and the corresponding weight coefficient w2 of different targets is marked according to the matching similarity;

[0079] 5) The weight coefficients w1 and w2 corresponding to different targets are multiplied to obtain the total possibility value w of different targets, and the total possibility value w is normalized and sorted, and the visible light camera and the gamma camera are marked and tracked on the fusion image.

[0080] In another implementation manner, in order to facilitate the detection of the region of view, the device is usually installed slightly inclined to the ground, so that the magnification of the image at different distances is different, which may cause inaccurate fusion effect and increased calculation error. The depth correction can be performed by selecting the features.

[0081] According to the spatial position of the device installation and the fusion registration relationship of the two, the depth corresponding relationship of the fusion display image and the three-dimensional space is obtained. For the device installed with a downward angle, when the feature points of different targets transmitted by the tracking algorithm module are obtained, the depth direction is corrected according to the area occupied by the face or the detected object features.

[0082] Y' = S / S d *Y index

[0083] Where S is the pixel area of ​​the visible light image occupied by the extracted features, S d To identify the area corresponding to the pixel position of the feature in the Y direction, Y index Y' represents the pixel position of the extracted feature point in the Y direction, and Y' represents the corrected pixel position of the feature point in the Y direction.

[0084] The following is about Figures 2-6 The process of merging and displaying data will be further explained. Figures 2-6 Originally a color image, it has been converted to grayscale for patent application purposes.

[0085] Figure 2 In the demonstration, a demonstrator holding radioactive material walks normally forward within the field of view. The tracking module extracts predetermined features of the target based on visible light images, specifically the demonstrator's facial features within the field of view. Based on the centroid of the facial features across multiple frames of visible light images within the field of view, the first trajectory of the target's predetermined feature points is determined. The gamma camera detects the radioactive material based on the count rate of received gamma photons, confirms the real-time framing frequency, reconstructs the radioactive image after recombining the gamma photon data, and determines the second trajectory of the radioactive material based on multiple frames of radioactive images within a predetermined time period. It is generally assumed that radioactive material is carried by a person or object, and the two trajectories show a certain similarity. Therefore, it is analyzed that synchronous post-processing analysis of the effective information in the visible light images can improve the accuracy of radioactive source localization.

[0086] Figures 2-6 The simulated test conditions used in the study were as follows: radionuclide: Tc99m, ~1.1mCi, dose equivalent rate at 4m ~0.15uSv / h, conventional background dose equivalent rate ~0.12uSv / h, and background count rate ~10kcps.

[0087] like Figures 3-5 As shown in the simulated test environment, a demonstrator walks through the visible light field at a normal speed. In this embodiment, the person with ID 4 on the left of the figure holds the radiation source behind them. The demonstrator walks for about 8 seconds, the alarm count rate threshold is set to ~15cps, and the continuous alarm time is about 7 seconds.

[0088] After the alarm, the system of the present application tracks and locates the radioactive source in real time, the light gray box is the fusion display area of the gamma camera on the visible light image, the blue points (the multiple rows of horizontal points on the body and near the body of the person in the figure) are the feature point distribution of different targets extracted, the green bottom number is the ID annotation of different targets, and the pseudo-color brightness is the probability degree of carrying radioactive material marked (the high-probability radioactive material position marked on the left side of ID4 personnel in the figure). That is, from Figures 3-5 As can be seen from the above, the system can track the person with ID4 suspected of carrying radioactive material.

[0089] During the continuous alarm time, the image of the radioactivity count is framed and reconstructed, the confidence sum of different targets is calculated according to the above process, and the confidence is calculated according to the above formula (1).

[0090] The results of this demonstration case are as follows:

[0091] ID 1 2 3 4 5 6 7 8 9 1s -182.88 -181.69 -Inf -175.87 -183.62 0.00 0.00 0.00 0.00 2s -199.67 -198.83 -Inf -194.85 -198.70 0.00 0.00 0.00 0.00 3s -294.71 -284.20 -Inf -268.76 -281.89 -294.15 0.00 0.00 0.00 4s -394.74 -369.82 -Inf -358.09 -369.38 -388.75 0.00 0.00 0.00 5s -Inf -466.22 -Inf -446.51 -456.45 -Inf 0.00 0.00 0.00 6s -Inf -392.14 -Inf -380.53 0.00 -Inf -Inf -Inf 0.00 7s -Inf -256.19 -Inf -250.52 0.00 0.00 -Inf -Inf -Inf Times 0 0 0 7 0 0 0 0 0

[0092] Explanation: ID4 is the person carrying the source in this test, ID1 and ID2 are the persons who do not carry the source and appear in the gamma image field of view at the same time, and the rest of the IDs are the false detections of the feature extraction algorithm. Confidence 0 represents that the ID has not appeared or disappeared on the corresponding frame image, and -Inf represents that the feature point is out of the effective area of the radiation image; Conclusion: As can be seen from the statistical data in the table, the relative frequency of ID4 is the highest for 7 times, so the possibility of carrying radioactive material is the highest, and it is consistent with the actual result.

[0093] Figure 6 As can be seen from the above, the system can track the person with ID4 suspected of carrying radioactive material.

[0094] Specifically, for the device installed in a downward angle, according to the spatial positions of the radioactive image acquisition device and the visible light image acquisition device and the fusion registration relationship between the two, the depth corresponding relationship of the fusion image of the radioactive image and the visible light image and the three-dimensional space is obtained, when the feature points of different targets are obtained by the tracking algorithm module, the depth direction of the face is corrected according to the area occupied by the face,

[0095] Y' = S / S d *Yindex

[0096] Wherein S is the pixel area of the visible light image occupied by the extracted face, S d is the standard face area corresponding to the Y direction pixel position of the identified human eye, Y index is the Y direction pixel position of the extracted feature point, and Y' is the corrected Y direction pixel position of the feature point.

[0097] Figure 6 The bifurcated line on the left side of the middle B is the intersection trajectory of the radiation source trajectory and the corrected radiation source trajectory, and the trajectory line on the right side is the face trajectory (the original image is a red trajectory). The left branch of the bifurcated line is the radiation source trajectory before correction (the original image is a green trajectory), and the right side is the radiation source trajectory after correction (the original image is a blue trajectory). It can be seen that the similarity between the face trajectory after correction and the radiation source trajectory is higher.

[0098] In the method of the present application, the confidence and the weight w1 of different targets are obtained from the radioactivity image at different time periods, and the similarity and the weight w2 of the radioactivity material trajectory and the trajectory of the target to be measured at the spatial position, so as to respectively strengthen the correlation between the radioactivity image distribution and the target to be measured in time and space, and further improve the accuracy of the screening of suspicious persons carrying radioactive materials.

[0099] Although the principles of the present application have been described in detail above in combination with the preferred embodiments of the present application, those skilled in the art should understand that the above embodiments are only illustrative implementations of the present application, and are not a limitation on the scope of the present application. The details in the embodiments do not constitute a limitation on the scope of the present application, and any equivalent transformation, simple replacement, etc. based on the technical solutions of the present application, which does not deviate from the spirit and scope of the present application, falls within the protection scope of the present application.

Claims

1. A method of detecting a target carrying a radioactive substance, characterized by, The method comprises: Step 1), collecting visible light images and radioactive images of the to-be-detected targets; Step 2), extracting predetermined features of each to-be-detected target based on the visible light images, and determining a first trajectory of the predetermined features; Step 3), determining a confidence sum value of the predetermined features of each to-be-detected target with respect to radioactivity, and assigning a first weight w1 to each to-be-detected target; Step 4), determining a second trajectory of radioactive substances based on the radioactive images, and determining a second weight w2 of each to-be-detected target according to the matching degree of the first trajectory and the second trajectory of each to-be-detected target; Step 5), determining the to-be-detected target carrying radioactive substances according to the first weight w1 and the second weight w2, The step 3) comprises mapping the predetermined features of each frame of visible light images to the radioactive images of the corresponding acquisition time along the first trajectory, determining the confidence sum value of the predetermined features of each to-be-detected target with respect to radioactivity, and assigning a first weight w1 to each to-be-detected target based on the confidence sum value of each to-be-detected target, The confidence sum value of the predetermined features with respect to radioactivity is determined by using the following formula: wherein p is the number of gamma photons detected by each detector pixel j of the radiological image acquisition device j the vector formed, j is the detector pixel index of the radiological image acquisition device, i is the pixel index after the gamma photon incident angle plane is discretized, f is the vector distribution of the number of gamma photons incident to the front surface of the radiological image acquisition device, f i is the number of gamma photons incident to the front surface of the radiological image acquisition device from the discretized target angle plane, c ij is the system transmission matrix quantitatively describing the response of the radiological image acquisition device to the incident gamma photons, the confidence sum value is positively correlated with the first weight w1, and the predetermined feature includes a face, a human body part, or a carried article.

2. The method of claim 1, wherein the radioactive material-carrying object is detected by using a radiation detector. The method further comprises obtaining a fusion image based on the visible light images and the radioactive images, labeling the to-be-detected target carrying radioactive substances in the fusion image based on the determined to-be-detected target, and further comprising collecting a radioactive count rate of the target region, issuing an alarm signal based on a radioactive count rate threshold, and performing the detection method in steps 1)-5) in response to the alarm signal.

3. A system for detecting a target carrying a radioactive substance, characterized in that The system comprises a visible light image acquisition device, a radioactive image acquisition device, a reconstruction module, a tracking module, a positioning module, and a discrimination module, The visible light image acquisition device is configured to acquire visible light images of the to-be-detected targets; The radioactive image acquisition device is configured to acquire a radioactive count rate of the to-be-detected targets; The reconstruction module is configured to reconstruct radioactive images based on the radioactive count rate; The tracking module is configured to extract predetermined features of each to-be-detected target based on the visible light images, and obtain a first trajectory based on the predetermined features; The positioning module is configured to determine a confidence sum value of the predetermined features of each to-be-detected target with respect to radioactivity, and assign a first weight w1 to each to-be-detected target based on the confidence sum value of each to-be-detected target; The discrimination module is configured to determine a second trajectory of radioactive substances based on a plurality of radioactive images output by the reconstruction module within a predetermined time, match the first trajectory and the second trajectory of each to-be-detected target, determine a second weight w2 of each to-be-detected target based on the matching similarity, and combine the first weight w1 and the second weight w2 corresponding to each to-be-detected target to obtain a confidence weight w of each to-be-detected target with respect to radioactivity, and determine the to-be-detected target with radioactive substances based on the weight ranking, The positioning module is configured to map the predetermined features of each frame of the visible light image to the radioactive image at a corresponding time along the first trajectory, determine a confidence sum value of the predetermined features with respect to radioactivity for each target to be detected, and assign a first weight w1 to each target to be detected based on the confidence sum value. The positioning module is configured to determine the confidence sum value of the predetermined features with respect to radioactivity according to the following formula: where p is the number of gamma photons detected by each detector pixel j of the radiological image acquisition device j the vector formed, j is the detector pixel index of the radiological image acquisition device, i is the pixel index after the gamma photon incidence angle plane is discretized, f is the vector distribution of the number of gamma photons incident to the front surface of the collimator of the radiological image acquisition device, c ij is the system transmission matrix for quantitatively describing the response of the radiological image acquisition device to the incident gamma photons.

4. The system for detecting a radioactive material carrying object according to claim 3, wherein The reconstruction module is further configured to construct a fusion image based on the visible light image and the radioactive count rate image, and the target detection system further comprises a display module configured to mark the target with radioactive material in the fusion image based on the determined position of the target.

5. The system for detecting a radioactive material carrying object according to claim 3, wherein The radioactive image acquisition device is a gamma camera, which recombines gamma photon data according to count rate information of received gamma photons and confirms a real-time framing frequency.

6. The system for detecting a radioactive material carrying object according to claim 3, wherein For the radioactive image acquisition device and the visible light image acquisition device installed in a downward inclined manner, the fusion image and the depth correspondence relationship in the three-dimensional space of the radioactive image and the visible light image are obtained according to the spatial positions of the radioactive image acquisition device and the visible light image acquisition device and the fusion registration relationship therebetween. When the feature points of different targets transmitted by the tracking module are obtained, the depth direction of the face is corrected according to the area occupied by the face, and the correction formula is as follows: Y' = S / Sd * Y index Sd is the standard face area corresponding to the Y direction pixel position of the recognized human eye, Y index Y is the Y direction pixel position of the extracted feature point, and Y' is the corrected Y direction pixel position of the feature point.

Images