Duplicate removal method and device for detection target

Abstract

The invention provides a duplicate removal method and device for a detection target. The method comprises the following steps: acquiring suspicious articles in human body images of the same scanning object in at least two different directions; defining a suspicious article located at the edge of a human body contour in a first human body image in the at least two human body images in different directions as a to-be-de-duplicated target; based on the target to be subjected to duplicate removal, performing duplicate removal on suspicious articles in a second human body image in a direction different from that of the first human body image in the at least two human body images in different directions to obtain a weight counting number of the suspicious articles in the second human body image; and determining the total number of the suspicious articles carried by the scanning object based on the weighing number and the number of the suspicious articles in the first human body image. The method can effectively identify and remove the same suspicious article which is repeatedly detected in the images in different directions, avoids repeated counting, and improves the security check efficiency.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method and device for deduplicating detected targets. Background Art

[0002] Millimeter wave human body imaging equipment is a security inspection instrument that uses millimeter wave technology. The equipment can efficiently detect suspicious items hidden under human clothing.

[0003] When millimeter-wave body imaging equipment is used in security inspections, if the device detects more than a certain number of suspicious items, staff will be prompted to conduct a full-body manual inspection of the passenger being inspected. During the identification and detection of suspicious items, some suspicious items may be identified multiple times in different body images, resulting in an increase in the number of suspicious items detected by the equipment, increasing the probability of staff conducting a full manual inspection of the passenger, and thus reducing security inspection efficiency. Summary of the Invention

[0004] The purpose of this application is to provide a method for deduplicating detection targets to solve the technical problem that the increase in the probability of full inspection leads to a decrease in security inspection efficiency.

[0005] In a first aspect, the present application provides a method for deduplicating a detection target, the method comprising: Obtain suspicious objects in human body images of the same scanned object in at least two different directions; defining a suspicious object located at a boundary of a human body contour in a first human body image of the at least two human body images in different directions as a target to be deduplicated; Deduplicating suspicious items in a second human body image in a different direction from the first human body image among the at least two human body images in different directions based on the target to be deduplicated, and obtaining a recounted number of suspicious items in the second human body image; The total number of suspicious items carried by the scanned object is determined based on the recounted number and the number of suspicious items in the first human body image.

[0006] Optionally, defining a suspicious object located at a boundary of a human body contour in a first human body image of the at least two human body images in different directions as a target to be deduplicated includes: When the total number of suspicious objects in the at least two human body images in different directions exceeds a re-inspection threshold, the suspicious objects located at the boundary of the human body contour in the first human body image of the at least two human body images in different directions are defined as targets to be deduplicated.

[0007] Optionally, defining a suspicious object located at a boundary of a human body contour in a first human body image of the at least two human body images in different directions as a target to be deduplicated includes: Using a human body image in at least one direction, determining a human body part mask map; The suspicious object in the first human body image has multiple boundary points, and the suspicious object in the first human body image with at least one boundary point located in the background area of ​​the human body part mask image is used as a target to be deduplicated; and / or Based on the human body part mask image, the human body contour boundary is determined, and suspicious objects in the first human body image whose distance to the human body contour boundary is less than the deduplication distance are taken as targets to be deduplicated.

[0008] Optionally, the suspicious objects in the first human body image whose distance to the human body outline boundary is less than a deduplication distance are taken as targets to be deduplicated, including: Determine the minimum distance between each boundary point of the suspicious object and each boundary point of the human body outline in the first human body image as the distance between each boundary point of the suspicious object and the human body outline; Determine the minimum distance between all boundary points of the suspicious object in the first human body image and the boundary of the human body outline as the distance from the suspicious object to the boundary of the human body outline; Based on the distance between each suspicious object in the first human image and the human body contour boundary, suspicious objects in the first human image whose distance to the human body contour boundary is less than the deduplication distance are taken as targets to be deduplicated.

[0009] Optionally, deduplicating suspicious items in a second human body image in a different direction from the first human body image among the at least two human body images in different directions based on the target to be deduplicated, and obtaining the recounted number of suspicious items in the second human body image includes: One of the first human body image and the second human body image is a front human body image, and the other is a back human body image; Mirroring the target to be deduplicated in the first human image horizontally to the second human image to obtain the target to be deduplicated after mirroring; wherein the vertical coordinates of the boundary points of the target to be deduplicated before mirroring are the same as the vertical coordinates of the boundary points of the target to be deduplicated after mirroring; Based on the mirrored target to be deduplicated, removing suspicious items in the second human body image that overlap with the mirrored target to be deduplicated and / or suspicious items whose center point is less than a desired distance from the center point of the mirrored target to be deduplicated; The number of suspicious objects remaining in the second human body image is used as the recounted number of suspicious objects in the second human body image.

[0010] Optionally, deduplicating suspicious items in a second human body image in a different direction from the first human body image among the at least two human body images in different directions based on the target to be deduplicated, and obtaining the recounted number of suspicious items in the second human body image includes: The first human body image is a front human body image and / or a back human body image, and the second human body image is a side human body image; Mapping the coordinates of the center position of the suspicious object in the first human body image into x- and y-coordinates in a world coordinate system, and mapping the coordinates of the center position of the human body part corresponding to the center position of the suspicious object into z-coordinates in the world coordinate system, to obtain first mapped coordinates of the suspicious object in the first human body image; Converting the first mapping coordinates into the second human body image to obtain second mapping coordinates; If the second mapping coordinates are located inside the suspicious object in the second human body image, deduplication is performed on the suspicious object.

[0011] Optionally, mapping the coordinates of the center position of the human body part corresponding to the center position of the suspicious object to a z coordinate in a world coordinate system includes: The z coordinate of the center point of the human body part on the same horizontal plane as the center position of the suspicious object is used as the z coordinate in the world coordinate system.

[0012] Optionally, using a human body image in at least one direction to determine a human body part mask map includes: Performing background segmentation on the human body image in at least one direction by using threshold segmentation, region segmentation, edge segmentation or clustering to obtain a background area and a human body target area; Extracting joints and / or skeletons from the target human body area, and clustering the extracted joints and / or skeletons; Dividing the human body target area according to the clustering results and the preset human body part labels to obtain areas corresponding to various human body parts that match the human body part labels; Pixel values ​​are set for the area corresponding to each divided human body part and the background area to obtain a human body part mask map.

[0013] In a second aspect, the present application further provides a device for detecting target deduplication, the device for detecting target deduplication comprising: processor; The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the method for deduplicating detection targets of the first aspect is implemented.

[0014] In a third aspect, the present application further provides a millimeter wave device, including a scanner, a processor, and a memory; The scanner is used to scan the scanning object; The processor is configured to define a suspicious item located at a boundary of a human body outline in a first human body image of the at least two human body images in different directions as a target to be deduplicated; to deduplicate suspicious items in a second human body image of the at least two human body images in a different direction from the first human body image based on the target to be deduplicated, to obtain a recounted number of suspicious items in the second human body image; and to determine a total number of suspicious items carried by the scanned object based on the recounted number and the number of suspicious items in the first human body image. The memory is used to store the total number of suspicious items carried by the scanned object.

[0015] The deduplication method for detecting targets provided in the embodiment of the present application deduplicates suspicious items located at the boundary of the human body contour in human body images in different directions, thereby avoiding repeated detection of suspicious items located at the boundary of the human body contour in human body images from another direction of view. Since only suspicious items close to the boundary of the human body contour are deduplicated instead of deduplicating all suspicious items, the computational complexity of the deduplication operation can be effectively reduced, and the deduplication process can be completed quickly. Moreover, based on the number of suspicious items in the first human body image and the recounted number of suspicious items in the second human body image, the total number of suspicious items carried by the scanned object is determined, so that the total number of suspicious items detected is closer to the total number of suspicious items actually carried by the scanned object, avoiding the problem of increased probability of manual full inspection due to repeated identification and counting of suspicious items, thereby improving security inspection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0017] Figure 1 A flowchart of a method for deduplicating detection targets provided in an embodiment of the present application; Figure 2 A schematic diagram of a human body part mask provided in an embodiment of the present application; Figure 3 A flow chart of a method for determining a target to be deduplicated provided in an embodiment of the present application; Figure 4 A flow chart of a method for deduplicating suspicious items provided in an embodiment of the present application; Figure 5This is a diagram of the internal structure of a device for detecting target deduplication provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The present application will be described in detail below in conjunction with the specific embodiments shown in the accompanying drawings, but these embodiments do not limit the present application. Structural, methodological, or functional changes made by ordinary technicians in this field based on these embodiments are included in the scope of protection of the present application.

[0019] When millimeter-wave devices are used in security inspections, if the device detects more than a certain number of suspicious items, it prompts the inspector to conduct a full-body manual inspection of the subject. If a suspicious item is identified once using both the front and back images of a person, and the two counts are added together, the number of detected suspicious items will increase, increasing the probability of a full manual inspection, which in turn reduces security inspection efficiency.

[0020] like Figure 1 As shown, the embodiment of the present application provides a method for deduplicating a detection target, which specifically includes the following steps: Step S101: obtaining suspicious objects in human body images of the same scan object in at least two different directions.

[0021] Images of the same scanned subject from different orientations can include, but are not limited to, frontal images, back images, and side images. Suspicious objects in human images can include knives, liquids, guns, ammunition, and so on. Suspicious object detection is performed on human images from various orientations using an intelligent recognition algorithm. Exemplary intelligent recognition algorithms include, but are not limited to, the YOLO (You Only Look Once) algorithm or the SSD (Single Shot MultiBox Detector) algorithm.

[0022] The detected suspicious object has a boundary (e.g., a rectangle) with multiple boundary points. For example, the boundary of the suspicious object has four boundary points, which can be expressed as Cn={Pn1, Pn2, Pn3, Pn4}, Pnm=P(Xnm, Ynm), P(Xnm, Ynm) is the image pixel coordinate, n is the nth suspicious object, m is the mth boundary point of the nth suspicious object, m∈{1,2,3,4}; the center point On of the suspicious object is: On=( ); Among them, (Xn1, Yn1), (Xn2, Yn2), (Xn3, Yn3) and (Xn4, Yn4) are the image pixel coordinates of the four boundary points of the nth suspicious object.

[0023] For example, suspicious items are detected in a frontal image of a person in the frontal direction to obtain a suspicious item set A, where set A = {Ca1, Ca2, Ca3, …, Caj}, i.e., j suspicious items are detected in the frontal image of the person in the frontal direction, and Caj represents the image pixel coordinates of the jth suspicious item. Suspicious items are detected in a back image of a person in the back direction to obtain a suspicious item set B, where set B = {Cb1, Cb2, Cb3, …, Cbk}, i.e., k suspicious items are detected in the back image of the person in the back direction, and Cbk represents the image pixel coordinates of the kth suspicious item.

[0024] Step S102: defining suspicious objects located at the boundary of the human body contour in the first human body image of at least two human body images in different directions as targets to be deduplicated.

[0025] Among them, the first human body image can be a frontal image of the human body in the front direction. Considering that when the suspicious item is located at the edge of the body of the inspected object (i.e., the scanned object), the suspicious item will be identified once in the human body images in different directions, so that the suspicious item located at the edge of the scanned object's body will be repeatedly identified and counted. Therefore, the suspicious items located at the boundary of the human body contour among the multiple suspicious items in the first human body image are taken as targets to be deduplicated.

[0026] Exemplarily, the target to be deduplicated may be a suspicious item in set A whose distance from the human body outline boundary is less than a set distance, that is, a suspicious item in set A whose center point is less than a set distance from the human body outline boundary is a suspicious item to be deduplicated.

[0027] Step S103: Deduplicating suspicious items in at least two human body images in different directions and in a second human body image in a different direction from the first human body image based on the target to be deduplicated, and obtaining the recounted number of suspicious items in the second human body image.

[0028] Exemplarily, deduplication of suspicious items in the second human image can be achieved, but is not limited to, by the following method: the target to be deduplicated (i.e., a suspicious item located at the boundary of the human body outline) in the first human image (e.g., a frontal image of a human body in the frontal direction) can be mapped to a second human image (e.g., a back image or a side image of a human body) from a different perspective than the first through spatial position mapping rules (e.g., a three-dimensional coordinate system conversion or horizontal mirror mapping). All suspicious items in the second human image are traversed. If a suspicious item in the second human image overlaps with the target to be deduplicated, it is determined that the suspicious item has been repeatedly identified and counted and needs to be deduplicated. The number of suspicious items in the second human image after deduplication is counted as the recounted number of suspicious items in the second human image, thereby preventing the same suspicious item carried by the scanned object from being counted repeatedly in images in different directions, thereby improving the accuracy and efficiency of security inspection results.

[0029] Step S104: determining the total number of suspicious items carried by the scanned object based on the recounted number and the number of suspicious items in the first human body image.

[0030] If the first image of a person is a frontal image, and the second image is a back image, the total number of suspicious items carried by the scanned subject is the sum of the number of suspicious items in the first image and the recounted number of suspicious items in the second image. For example, if there are j suspicious items in the first image and k suspicious items in the second image, and the recounted number is r, then the recounted number of suspicious items in the second image is kr. The total number of suspicious items carried by the scanned subject, t, is: t = j + kr, where j, k, and r are all integers.

[0031] In an embodiment of the present application, suspicious items in human body images of the same scanned object in different directions are determined, and suspicious items located at the boundary of the human body contour in the first human body image are determined as targets to be deduplicated. All suspicious items in the second human body image with a direction different from that of the first human body image are traversed, and the suspicious items counted repeatedly in the second human body image are determined through the targets to be deduplicated, that is, the suspicious items that need to be deduplicated. Since these suspicious items located at the boundary of the human body contour will be repeatedly detected in the human body image from another direction of view, by deduplicating only the suspicious items close to the boundary of the human body contour (instead of all suspicious items), the computational amount of the deduplication operation can be reduced, and the deduplication process can be completed quickly.

[0032] In an optional embodiment of the present application, after step S104, a prompt may be issued when the total quantity t after recounting exceeds a set alarm threshold. The prompt may take the form of a light, a text reminder on a display screen, or an audio prompt, etc., without limitation. The prompt information may include, but is not limited to: "Please proceed to manual security inspection," "Special items exceeded," a light, a ring tone, etc.

[0033] The number of suspicious items after deduplication in the second human body image is counted as the recounted number of suspicious items in the second human body image, thereby preventing the same suspicious item located at the edge of the human body from being counted repeatedly in human body images in different directions; based on the number of suspicious items in the first human body image and the recounted number of suspicious items in the second human body image, the total number of suspicious items carried by the scanned object is determined, so that the total number of suspicious items detected is closer to the total number of suspicious items actually carried by the scanned object, avoiding the problem of increased probability of manual full inspection due to repeated identification and counting of suspicious items, thereby improving security inspection efficiency.

[0034] In an optional embodiment of the present application, before step S102, it can be determined whether the total number of suspicious items in at least two human body images in different directions exceeds the re-inspection threshold. When the total number exceeds the reconstruction threshold, step S102 is executed, that is, the suspicious items located at the edge of the human body contour in the first human body image in at least two human body images in different directions are defined as targets to be deduplicated. In this case, the number of deduplication times can be reduced.

[0035] Specifically, deduplication is triggered when the total number of suspicious items detected in at least two images of the same scanned subject from different orientations (e.g., front and back) (i.e., the sum of the number of suspicious items in each image before deduplication) exceeds a preset recheck threshold. If the total number of suspicious items detected does not reach the recheck threshold, deduplication is not required. This is because manual full inspection will not be triggered and the probability of manual full inspection will not be increased if the total number of suspicious items detected does not reach the recheck threshold. Therefore, deduplication is not performed on suspicious items if the total number of suspicious items detected does not reach the recheck threshold. This improves the efficiency of suspicious item inspections while minimizing the computational overhead.

[0036] In one embodiment, in step S102, suspicious objects located at the boundary of the human body outline in the first human body image of at least two human body images in different directions are defined as targets to be deduplicated, which can be achieved by, but is not limited to, the following steps: Step S201: using a human body image in at least one direction, determining a human body part mask image.

[0037] Among them, the size of the human body part mask map is the same as the size of the human body image. The human body part mask map is provided with pixel values ​​of different sizes. The pixel value is used to represent the area to which the current pixel point belongs. The area to which the current pixel point belongs is one of the areas corresponding to each human body part and the background area.

[0038] Exemplarily, as shown in FIG2 , a frontal image of a human body in the frontal direction is used to determine a human body part mask image. Figure 2 In the figure, a is a frontal image of a human body, in which two suspicious objects are currently detected. Figure 2 In the figure, b represents the human body part mask. As shown in Figure b, the pixel values ​​of different sizes in the human body part mask, for example, 0 to 14, are used to indicate the region to which the current pixel belongs, for example, whether it belongs to the region corresponding to the human body part or the background region. The human body parts corresponding to the pixel values ​​of different sizes are as follows: 1-head, 2-torso, 3-left upper arm, 4-right upper arm, 5-left lower arm, 6-right lower arm, 7-left hand, 8-right hand, 9-left thigh, 10-right thigh, 11-left calf, 12-right calf, 13-left leg, 14-right foot. The background region is 0.

[0039] Step S202: The suspicious object in the first human body image has multiple boundary points. The suspicious object in the first human body image with at least one boundary point located in the background area of ​​the human body part mask image is taken as a target to be deduplicated.

[0040] Assuming that the front image is used as the first human body image, for the suspicious item carried by the scanned object in the front image, if at least one boundary point of the suspicious item is not located in the area corresponding to the human body part in the human body part mask image, but is in the background area of ​​the human body part mask image, then it is determined that the suspicious item is located at the boundary of the human body contour and needs to be deduplicated. In this way, by determining that at least one boundary point of the suspicious item is located in the background area to determine the target to be deduplicated, the target to be deduplicated can be quickly located from multiple suspicious items in the first human body image simply and accurately.

[0041] In an embodiment of the present application, in addition to finding the target to be deduplicated from multiple suspicious items in the first human image through step S202, step S203 can also be used, or step S202 and step S203 can be combined to find the target to be deduplicated from multiple suspicious items in the first human image.

[0042] Step S203: Based on the human body part mask image, determine the human body contour boundary, and take the suspicious objects in the first human body image whose distance to the human body contour boundary is less than the deduplication distance as the target to be deduplicated.

[0043] For example, the human body contour boundary can be determined based on the human body part regions and the background region in the human body part mask image. For example, the regions corresponding to the human body parts in the human body part mask image are combined, and the human body contour boundary is determined based on the combined regions and the background region.

[0044] For another example, a foreground-background segmentation algorithm may be used to determine the boundaries of each human body part region (foreground) in the human body part mask image as the human body contour boundary.

[0045] For example, if the suspicious object's outline is a cuboid, the deduplication distance can be one-third of the suspicious object's shortest edge, or one-half of its longest edge. If the suspicious object's outline is circular, the deduplication distance can be one-fifth of the suspicious object's diameter. The deduplication distance can be an empirical value or configured and adjusted by the user through the interactive interface.

[0046] If the distance between the suspicious item carried by the scanned object and the boundary of the human body contour in the human body part mask image is less than the deduplication distance, it can be inferred that the suspicious item is a suspicious item located at the boundary of the human body contour, and repeated identification and counting will occur. Therefore, it needs to be treated as a target to be deduplicated.

[0047] In one embodiment, determining a human body part mask using a human body image in at least one direction may include: Performing background segmentation on the human body image in at least one direction by using threshold segmentation, region segmentation, edge segmentation or clustering to obtain a background area and a human body target area; Extracting joints and / or skeletons from the target area of ​​the human body, and clustering the extracted joints and / or skeletons; The human body target area is divided according to the clustering results and the preset human body part labels to obtain the area corresponding to each human body part that matches the human body part label; Pixel values ​​are set for the area corresponding to each divided human body part and the background area to obtain a human body part mask map.

[0048] Specifically, threshold segmentation, region segmentation, edge segmentation or clustering methods are used to perform background segmentation on a human body image in at least one direction (such as a front image of a human body or a back image of a human body), and the image is divided into a background area and a human body target area; joints and / or skeletons are extracted from the human body target area to identify key structural features of the human body, and cluster analysis is performed on the extracted joints and / or skeletons to determine the relative positional relationship of various parts of the human body; based on the clustering results and preset human body part labels (such as head, torso, limbs, etc.), the human body target area is accurately divided to obtain areas corresponding to various human body parts that match the preset labels; pixel values ​​are assigned to the areas corresponding to each divided human body part and the background area, thereby generating a human body part mask map that clearly identifies the areas corresponding to each human body part and the background area.

[0049] In one embodiment of the present application, Figure 3 As shown, in step S203, suspicious objects whose distance to the human body contour boundary in the first human body image is less than the deduplication distance are taken as targets to be deduplicated, which can be implemented as follows: Step S301: determining the minimum distance between each boundary point of the suspicious object and each boundary point of the human body contour boundary in the first human body image, as the distance between each boundary point of the suspicious object and the human body contour boundary.

[0050] In S301 , the distance between each boundary point of the suspicious object in the first human body image and each boundary point of the human body contour boundary may be determined, and the minimum distance therebetween may be used as the distance between each boundary point and the human body contour boundary.

[0051] Exemplarily, the boundary of a suspicious object is formed by four points in a geometric shape (e.g., a rectangle). The boundary point set for the nth suspicious object Cn in the first human image is {Pn1, Pn2, Pn3, Pn4}, where Pnm = P(Xnm, Ynm), where P(Xnm, Ynm) represents the image pixel coordinates, n represents the nth suspicious object, and m represents the mth boundary point of the nth suspicious object, where m∈{1,2,3,4}. The human body contour boundary has several boundary points, and Z represents a set of boundary points, Z={Pz1, Pz2, Pz3, …}. Exemplarily, Pz1 represents the image pixel coordinates of the first boundary point of the human body contour boundary.

[0052] For each boundary point of the nth suspicious object, for example, Pn1, calculate the Euclidean distance between it and each boundary point on the human body contour (Pz1, Pz2, Pz3, ...). This yields a set of distances from boundary point Pn1 to each boundary point on the human body contour (Pz1, Pz2, Pz3, ...). The minimum value from this set of distances is taken as the distance from boundary point Pn1 to the human body contour. Similarly, the distance from each boundary point of the nth suspicious object (Pn1, Pn2, Pn3, and Pn4) to the human body contour can be obtained.

[0053] Step S302: Determine the minimum distance between all boundary points of the suspicious object in the first human body image and the human body contour boundary as the distance from the suspicious object to the human body contour boundary.

[0054] For example, for the nth suspicious object, based on the distances from all boundary points (Pn1, Pn2, Pn3 and Pn4) of the nth suspicious object to the human body contour boundary, the minimum distance is taken as the distance from the nth suspicious object to the human body contour boundary, where the distance from the suspicious object to the human body contour boundary can be recorded as the distance Lcz.

[0055] Step S303: Based on the distance between each suspicious object and the human body contour in the first human body image, the suspicious objects in the first human body image whose distance to the human body contour is less than the deduplication distance are taken as targets to be deduplicated.

[0056] Specifically, based on the distances from n suspicious objects to the body boundary in the first human image, ie, n distances Lcz, a distance Lcz smaller than the deduplication distance is determined from the n distances Lcz, and the corresponding suspicious object is marked as a target to be deduplicated.

[0057] In an embodiment of the present application, the minimum distance between each suspicious object in the first human body image and the boundary of the human body contour is calculated to accurately identify suspicious objects located at the edge of the human body in the first human body image. These suspicious objects will be repeatedly detected in the human body image from another viewing angle. These suspicious objects in the first human body image are used as targets to be deduplicated, which can effectively reduce the computational complexity of the deduplication operation.

[0058] In one embodiment, Figure 4 As shown, in step S103, based on the target to be deduplicated, duplicate suspicious items are deduplicated in at least two human body images in different directions and in a second human body image in a direction different from the first human body image, to obtain the recounted number of suspicious items in the second human body image, which specifically includes the following steps: Step S401: one of the first human body image and the second human body image is a human body front image, and the other is a human body back image.

[0059] Step S402: horizontally mirroring the target to be deduplicated in the first human image to the second human image to obtain the target to be deduplicated after mirroring; wherein the vertical coordinates of the boundary points of the target to be deduplicated before mirroring are the same as the vertical coordinates of the boundary points of the target to be deduplicated after mirroring; Specifically, horizontal mirroring refers to transforming each boundary point of the object to be deduplicated into a second image symmetrically with the vertical midline of the first image as the axis of symmetry. The boundary points of the object to be deduplicated before and after the mirroring must meet the following conditions: the vertical coordinates of the boundary points of the object to be deduplicated before and after the mirroring remain unchanged; and the sum of the horizontal coordinates of the boundary points of the object to be deduplicated before and after the mirroring is equal to the width of the first image.

[0060] For example, assuming that Pa (Xa, Ya) represents the pixel coordinates of the boundary point to be deduplicated before mirroring, Pb (Xb, Yb) represents the pixel coordinates of the boundary point to be deduplicated after mirroring, and the width of the first human body image is W, then its coordinates have the following relationship: Xb = W-Xa, Yb = Ya.

[0061] Step S403: Based on the mirrored target to be deduplicated, remove suspicious items in the second human image that overlap with the mirrored target to be deduplicated and / or suspicious items whose center point is less than a desired distance from the center point of the mirrored target to be deduplicated; Specifically, all suspicious items in the second human image are compared one by one against the mirrored target to be deduplicated, to determine whether any suspicious items overlap with or are too close to the mirrored target. A suspicious item that overlaps with the mirrored target is considered to overlap if, after the mirroring operation, the overlapping area of ​​two or more suspicious items exceeds an area overlap threshold (e.g., 90% or 80%, depending on the detection requirements).

[0062] Exemplarily, if a suspicious object in the second human body image has an overlapping area greater than a certain area with the boundary of the target to be deduplicated after mirroring, or the distance between the center point of the suspicious object and the center point of the target to be deduplicated after mirroring is less than a preset expected distance threshold, then it is considered that the two targets are actually repeated detections of the same suspicious object in images in different directions, and the suspicious objects that meet the above conditions in the second human body image are deduplicated, thereby avoiding the same suspicious object located at the edge of the human body from being counted repeatedly in human body images in different directions. Further optionally, if there are suspicious objects that coincide with the position of the target to be deduplicated after mirroring, and are too close to the target to be deduplicated after mirroring, then these suspicious objects can be simultaneously determined as suspicious objects of repeated detection. The values ​​of the various thresholds mentioned in the embodiments of the present application can be flexibly configured according to the detection requirements.

[0063] Step S404: taking the number of suspicious objects remaining in the second human body image as the recounted number of suspicious objects in the second human body image.

[0064] In an embodiment of the present application, the target to be deduplicated in the front image of the human body is mapped to the corresponding position of the back image of the human body through mirror mapping, or the target to be deduplicated in the back image of the human body is mapped to the corresponding position of the front image of the human body, thereby performing matching and deduplication. This can simply and efficiently solve the problem of repeated counting of the same suspicious object located at the edge of the human body under multi-directional viewing angles.

[0065] In one embodiment, deduplicating suspicious items in a second human body image in a different orientation from the first human body image in at least two human body images in different orientations based on a target to be deduplicated, and obtaining a recounted number of suspicious items in the second human body image includes: The first human body image is a front image of a human body and / or a back image of a human body, and the second human body image is a side image of a human body; Mapping the coordinates of the center position of the suspicious object in the first human body image into x- and y-coordinates in a world coordinate system, and mapping the coordinates of the center position of the human body part corresponding to the center position of the suspicious object into z-coordinates in the world coordinate system, to obtain first mapped coordinates of the suspicious object in the first human body image; Converting the first mapping coordinates into the second human body image to obtain second mapping coordinates; If the second mapping coordinates are located inside the suspicious object in the second human body image, the suspicious object is deduplicated.

[0066] Exemplarily, the first human image can be a frontal or back image of the human body, and the second human image can be a side image of the human body. The suspicious item identified in the first human image is processed by mapping the image coordinates of the center of the suspicious item to the world coordinate system, corresponding to the x and y coordinates of the world coordinate system, respectively. The image coordinates of the center of the body part where the suspicious item is located are mapped to the z coordinate of the world coordinate system, thereby obtaining the first mapped coordinates (x, y, z) of the suspicious item in the world coordinate system. Based on the coordinate transformation relationship between the world coordinate system and the second human image, the first mapped coordinates (x, y, z) are transformed into the image coordinate system of the second human image to obtain the second mapped coordinates. A determination is then made as to whether the second mapped coordinates lie within the bounding box of a suspicious item in the second human image. If the second mapped coordinates lie within the bounding box of a suspicious item, it indicates that the suspicious item and the suspicious item in the first human image are spatially overlapping or highly similar, indicating duplicate detection of the same item. Therefore, the suspicious item in the second human image is deduplicated. In this embodiment of the present application, this method effectively prevents the same suspicious item from being counted repeatedly in images from different perspectives, thereby improving the accuracy and reliability of security inspection results.

[0067] In one embodiment, mapping the coordinates of the center position of the human body part corresponding to the center position of the suspicious object to the z coordinate in the world coordinate system specifically includes: The z coordinate of the center point of the human body part on the same horizontal plane as the center position of the suspicious object is used as the z coordinate in the world coordinate system.

[0068] For example, to determine the spatial position of a suspicious object in a world coordinate system, the center of the suspicious object in a first human image (e.g., a frontal image or a back image) is first identified, and its image coordinates are mapped to x- and y-coordinates in the world coordinate system. Using the horizontal plane where the center of the suspicious object lies, the center point of a human body part on that plane (e.g., the horizontal center point of a human body part such as the waist, chest, or abdomen) is found, and the z-coordinate of this center point is output as the z-coordinate in the world coordinate system. In this embodiment of the present application, using human body parts as a reference when determining the spatial position of a suspicious object can reduce fluctuations in the estimated position of the suspicious object.

[0069] In an optional embodiment of the present application, after the location of suspicious items is estimated and duplicates are removed, a prompt can be given when the total number t after the recount is greater than a set alarm threshold. The prompt can be in the form of a light, a text reminder on the display screen, or an audio prompt, which is not limited here. The prompt information can include, but is not limited to: please conduct manual security inspection, special items are exceeded, light or ring tone, etc. In this way, the prompt is only given when the total number of suspicious items after duplicate removal is greater than the alarm threshold, which can reduce the false alarm rate and the frequency of manual security inspections, thereby improving security inspection efficiency.

[0070] Based on the same concept, an embodiment of the present application further provides a millimeter wave device, including a scanner, a processor, and a memory; A scanner, used for scanning an object; The processor is configured to define a suspicious item located at a boundary of a human body outline in a first human body image of at least two human body images in different directions as a target to be deduplicated; to deduplicate suspicious items in a second human body image of at least two human body images in different directions, which is in a different direction from the first human body image, based on the target to be deduplicated, to obtain a recounted number of suspicious items in the second human body image; and to determine a total number of suspicious items carried by the scanned object based on the recounted number and the number of suspicious items in the first human body image. The memory is used to store the total number of suspicious items carried by the scanned object.

[0071] Based on the same inventive concept, the embodiment of the present application also provides a device for detecting target deduplication. Figure 5 Schematic diagram of the structure of the device for detecting target deduplication provided by the embodiment of the present application, such as Figure 5 As shown, illustratively, the device 50 for detecting target deduplication may include a first processor 51 .

[0072] Exemplarily, the device 50 for detecting target deduplication may further include a memory 52 and a transceiver 53 .

[0073] The first processor 51 , the memory 52 , and the transceiver 53 may be connected via a communication bus.

[0074] The following combination Figure 5 The example of FIG. 5 introduces various components of the device 50 for detecting target deduplication: The device 50 for detecting target deduplication may include the following components: a first processor 51 may be a single processor or a collective term for multiple processing elements. For example, the first processor 51 may be one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present application, such as one or more microprocessors (digital signal processors, DSPs) or one or more field programmable gate arrays (FPGAs).

[0075] Exemplarily, the first processor 51 may execute various functions of the device 50 for detecting target deduplication by running or executing a software program stored in the memory 52 and calling values ​​stored in the memory 52 .

[0076] In a specific implementation, as an embodiment, the first processor 51 may include one or more CPUs, such as Figure 5 CPU0 and CPU1 are shown in FIG.

[0077] As an optional embodiment, the device 50 for detecting target deduplication may also include multiple processors, such as Figure 5 , a first processor 51 and a second processor 54 are shown. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). A processor herein can refer to one or more devices, circuits, and / or processing cores for processing values ​​(e.g., computer program instructions).

[0078] The memory 52 is used to store the software program for executing the solution of the present invention, and is controlled by the first processor 51 to execute. The specific implementation method can refer to the above-mentioned detection target deduplication method embodiment, which will not be repeated here.

[0079] Exemplarily, the memory 52 may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compact disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of an instruction or value structure and can be accessed by a computer, but is not limited thereto. The memory 52 may be integrated with the first processor 51 or exist independently and accessed through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 51, which is not specifically limited in this embodiment of the present application.

[0080] The transceiver 53 is used to communicate with the network device or the terminal device.

[0081] For example, the transceiver 53 may include a receiver and a transmitter ( Figure 5 The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0082] For example, the transceiver 53 may be integrated with the first processor 51 or may exist independently and communicate with the first processor 51 through the interface circuit ( Figure 5 (not shown) is coupled to the first processor 51, which is not specifically limited in this embodiment of the present application.

[0083] Exemplarily, the device 50 for detecting target deduplication may also include a display 55 for displaying the rendering result. The display 55 may be a liquid crystal display or an electronic ink display, which is not specifically limited in the embodiment of the present application.

[0084] In addition, the technical effects of the device 50 for detecting target deduplication can refer to the technical effects of the method for detecting target deduplication in the above method embodiment, and will not be repeated here.

[0085] Based on the same inventive concept, an embodiment of the present application also provides an electronic device, comprising: at least one memory and at least one processor, the at least one memory storing executable code, and the at least one processor being used to execute the executable code in the at least one memory to implement the above-mentioned detection target deduplication method.

[0086] It should be understood that the processor in the embodiments of the present application may be a central processing unit (CPU), but may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0087] It should also be understood that the memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0088] The above embodiments can be implemented in whole or in part via software, hardware (e.g., circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. A computer program product comprises one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the processes or functions according to the embodiments of the present application are fully or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or value center to another website, computer, server, or value center via infrared, microwave, or other means. A computer-readable storage medium can be any available medium accessible by a computer, or a storage device such as a server or value center that contains a collection of one or more available media. Available media can include magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), solid-state drives, etc.

[0089] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0090] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0091] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0092] Each embodiment in this specification is described in a related manner. Similar parts between the embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences from other embodiments. In particular, the device embodiments are generally similar to the method embodiments, so the description is relatively simple. For related parts, refer to the description of the method embodiments.

[0093] The above are only preferred embodiments of the present application and are not intended to limit the scope of protection of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application are included in the scope of protection of the present application.

Claims

1. A method for deduplicating a detection target, characterized in that: The method comprises: Obtain suspicious objects in human body images of the same scanned object in at least two different directions; defining a suspicious object located at a boundary of a human body contour in a first human body image of the at least two human body images in different directions as a target to be deduplicated; Deduplicating suspicious items in a second human body image in a different direction from the first human body image among the at least two human body images in different directions based on the target to be deduplicated, and obtaining a recounted number of suspicious items in the second human body image; The total number of suspicious items carried by the scanned object is determined based on the recounted number and the number of suspicious items in the first human body image.

2. The method for deduplicating a detection target according to claim 1, wherein: Defining a suspicious object located at the boundary of the human body outline in the first human body image of the at least two human body images in different directions as a target to be deduplicated includes: When the total number of suspicious objects in the at least two human body images in different directions exceeds a re-inspection threshold, the suspicious objects located at the boundary of the human body contour in the first human body image of the at least two human body images in different directions are defined as targets to be deduplicated.

3. The method for deduplicating a detection target according to claim 1, wherein: Defining a suspicious object located at the boundary of the human body contour in the first human body image of the at least two human body images in different directions as a target to be deduplicated includes: Using a human body image in at least one direction, determining a human body part mask map; The suspicious object in the first human body image has multiple boundary points, and the suspicious object in the first human body image with at least one boundary point located in the background area of ​​the human body part mask image is used as a target to be deduplicated; and / or Based on the human body part mask image, the human body contour boundary is determined, and suspicious objects in the first human body image whose distance to the human body contour boundary is less than the deduplication distance are taken as targets to be deduplicated.

4. The method for deduplicating a detection target according to claim 3, wherein: The method further comprises: determining, in the first human body image, suspicious objects whose distance from the human body contour boundary is less than a deduplication distance as targets to be deduplicated, including: Determine the minimum distance between each boundary point of the suspicious object and each boundary point of the human body outline in the first human body image as the distance between each boundary point of the suspicious object and the human body outline; Determine the minimum distance between all boundary points of the suspicious object in the first human body image and the boundary of the human body outline as the distance from the suspicious object to the boundary of the human body outline; Based on the distance between each suspicious object in the first human image and the human body contour boundary, suspicious objects in the first human image whose distance to the human body contour boundary is less than the deduplication distance are taken as targets to be deduplicated.

5. The method for deduplicating a detection target according to claim 1, wherein: Deduplicating suspicious items in a second human body image in a different direction from the first human body image among the at least two human body images in different directions based on the target to be deduplicated, and obtaining a recounted number of suspicious items in the second human body image, comprising: One of the first human body image and the second human body image is a front human body image, and the other is a back human body image; Mirroring the target to be deduplicated in the first human image horizontally to the second human image to obtain the target to be deduplicated after mirroring; wherein the vertical coordinates of the boundary points of the target to be deduplicated before mirroring are the same as the vertical coordinates of the boundary points of the target to be deduplicated after mirroring; Based on the mirrored target to be deduplicated, removing suspicious items in the second human body image that overlap with the mirrored target to be deduplicated and / or suspicious items whose center point is less than a desired distance from the center point of the mirrored target to be deduplicated; The number of suspicious objects remaining in the second human body image is used as the recounted number of suspicious objects in the second human body image.

6. The method for deduplicating a detection target according to claim 1, wherein: Deduplicating suspicious items in a second human body image in a different direction from the first human body image among the at least two human body images in different directions based on the target to be deduplicated, and obtaining a recounted number of suspicious items in the second human body image, comprising: The first human body image is a front human body image and / or a back human body image, and the second human body image is a side human body image; Mapping the coordinates of the center position of the suspicious object in the first human body image into x- and y-coordinates in a world coordinate system, and mapping the coordinates of the center position of the human body part corresponding to the center position of the suspicious object into z-coordinates in the world coordinate system, to obtain first mapped coordinates of the suspicious object in the first human body image; Converting the first mapping coordinates into the second human body image to obtain second mapping coordinates; If the second mapping coordinates are located inside the suspicious object in the second human body image, deduplication is performed on the suspicious object.

7. The method for deduplicating a detection target according to claim 6, wherein: Mapping the coordinates of the center position of the human body part corresponding to the center position of the suspicious object to the z coordinate in the world coordinate system includes: The z coordinate of the center point of the human body part on the same horizontal plane as the center position of the suspicious object is used as the z coordinate in the world coordinate system.

8. The method for deduplicating a detection target according to claim 3, wherein: Using a human body image in at least one direction, determining a human body part mask map, including: Performing background segmentation on the human body image in at least one direction by using threshold segmentation, region segmentation, edge segmentation or clustering to obtain a background area and a human body target area; Extracting joints and / or skeletons from the target human body area, and clustering the extracted joints and / or skeletons; Dividing the human body target area according to the clustering results and the preset human body part labels to obtain areas corresponding to various human body parts that match the human body part labels; Pixel values ​​are set for the area corresponding to each divided human body part and the background area to obtain a human body part mask map.

9. A device for detecting target deduplication, characterized in that: The device for detecting target deduplication includes: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method for deduplication of detection targets according to any one of claims 1 to 8 is implemented.

10. A millimeter wave device, characterized in that: Includes scanner, processor and memory; The scanner is used to scan the scanning object; The processor is configured to define a suspicious object located at a boundary of a human body outline in a first human body image of the at least two human body images in different directions as a target to be deduplicated; for deduplicating suspicious items in the at least two human body images in different directions and in a second human body image in a different direction from the first human body image based on the target to be deduplicated, to obtain a recounted number of suspicious items in the second human body image; and for determining the total number of suspicious items carried by the scanned object based on the recounted number and the number of suspicious items in the first human body image; The memory is used to store the total number of suspicious items carried by the scanned object.

Images