Millimeter wave image target mapping method, device, equipment, medium and product

Abstract

The invention discloses a millimeter wave image target mapping method, device and equipment, a medium and a product. The method comprises the following steps: firstly, calculating a central point position of a non-arm part target frame in a doll graph by adopting a non-arm part mapping strategy, and carrying out zooming adjustment on the non-arm part target frame; secondly, an arm part mapping strategy is adopted, the center point position of an arm part target frame in the doll graph is calculated, and zooming adjustment is conducted on the arm part target frame; and finally, mapping the adjusted non-arm part target frame and the adjusted arm part target frame into a standard doll graph based on the center point of the target frame in the obtained doll graph. According to the method, different processing strategies are respectively designed for the non-arm part target frame and the arm part target frame, so that the mapping accuracy and the visual consistency are improved.

Description

Technical Field

[0001] This invention belongs to the field of human body millimeter-wave image target detection, specifically relating to a millimeter-wave image target mapping method, device, equipment, medium, and product. Background Technology

[0002] Millimeter-wave image target detection is crucial for detecting contraband carried on the human body and can be widely applied in security checks at airports, train stations, and other locations, serving as an effective alternative to existing human body screening methods. Because millimeter-wave images involve sensitive human information, target detection results cannot be simply labeled and displayed directly on the original image. Currently, the widely accepted and adopted method is to map the target detection results from the original millimeter-wave image onto a standard mannequin image before displaying them.

[0003] The current strategy for target mapping is to combine the center point of the target bounding box with the associated human features to draw the position of the target bounding box on the mannequin image. However, due to the huge differences in shape and posture between the mannequin and the natural person being detected, positional offsets and errors in target mapping often occur, which seriously affect the performance and practical use of the detection equipment.

[0004] In real-world applications, different body parts of a person being detected, such as non-arm parts (including the torso and legs) and arm parts (including the upper arm and forearm), exhibit significant differences in morphological changes and posture: non-arm parts are relatively stable in a normal standing posture, so the relative positions of various body parts within this area remain stable; however, arm parts have greater postural flexibility, making their relative positions within this area, as well as their relative positions to parts outside this area, more prone to change. Based on these differences, the mapping process for non-arm and arm parts presents different challenges. For targets outside the arms, the differences in shape and posture between the human body in different detected natural person images and the puppet in the standard puppet image only cause changes in the relative position and size of the target in the detected natural person image and the puppet image, thus causing mapping deviation. However, for targets in the arms, the arm postures of the detected natural person vary greatly, and the spread angles are different, which is very different from the fixed arm posture of the standard puppet. This difference can easily cause the arm targets and their adjacent targets to have shape changes, misalignments, and misplacements during the mapping process, resulting in mapping errors. Summary of the Invention

[0005] To address the different challenges in mapping targets in non-arm and arm regions, this invention provides a millimeter-wave image target mapping method, apparatus, device, medium, and product. The mapping method employs different processing strategies for non-arm and arm region target bounding boxes, aiming to improve mapping accuracy and visual consistency.

[0006] In a first aspect, embodiments of this application provide a millimeter-wave image target mapping method, comprising the following steps: The Cartesian coordinate system described below is established with the top-left corner of the image as the origin, where the origin points to the right and downwards respectively. shaft and The positive direction of the axis. It should be noted that the human keypoints and bounding boxes in the millimeter-wave images shown below are obtained using methods such as human pose estimation to describe human pose and structure. These keypoints include, but are not limited to, the positions of the head, shoulders, elbows, wrists, torso, hips, knees, and ankles.

[0007] First, a non-arm part mapping strategy is adopted. Based on the relative positional relationship between corresponding elements in the original image and the puppet image in the vertical and horizontal directions, the center point position of the target box of the non-arm part in the puppet image is calculated, and the target box of the non-arm part is scaled and adjusted.

[0008] Secondly, an arm part mapping strategy is adopted. Based on the relationship between the distance in the original image and the distance between the corresponding points in the puppet image, as well as the principle of similar triangles, the center point position of the target box of the arm part in the puppet image is calculated, and the target box of the arm part is scaled and adjusted.

[0009] Finally, based on the center point of the target bounding box in the obtained puppet image, the adjusted non-arm part target bounding box and arm part target bounding box are mapped onto the standard puppet image.

[0010] In one possible implementation, the non-arm area mapping strategy is as follows: In the original image, the center point of the target bounding box (excluding the arm area), the human body's central axis, and the two closest human body keypoints perpendicularly to the same side of the target bounding box are proportionally related to the center point of the corresponding target bounding box, the human body's central axis, and the two corresponding human body keypoints in the puppet image. Using these existing distances, the center point of the target bounding box (excluding the arm area) in the puppet image is calculated. Then, the target bounding box is scaled and adjusted based on the aspect ratio of the human body bounding box in the puppet image and the original image.

[0011] In one possible implementation, the arm region mapping strategy is as follows: The distances between the two closest vertical arm keypoints on the same side of the target bounding box in the original image, and the distance between one of these keypoints and the center point of the target bounding box, are proportional to the distances between corresponding points in the puppet image. Furthermore, based on the principle of similar triangles, the difference between the angle between the line connecting the two closest vertical arm keypoints on the same side of the target bounding box in the original image and the horizontal direction, and the angle between the line connecting one of these keypoints and the center point of the target bounding box, remains unchanged after mapping to the puppet image. Using these existing conditions, the center point position of the target bounding box for the arm portion in the puppet image is calculated. Then, the target bounding box is scaled and adjusted based on the aspect ratio of the human body bounding box in the puppet image and the original image.

[0012] In one possible implementation, non-arm part mapping is performed using a non-arm part mapping strategy, specifically including the following steps: Step 1: Calculate the horizontal coordinate of the human body's central axis based on the location of key points on the human body, and determine whether the target box is located on the left or right side of the human body; in addition, if the horizontal coordinates are equal, it is determined that the target box is located on the left side of the human body.

[0013] Step 2: Select the center point of the target bounding box as the reference point P0. Calculate the vertical distances between the reference point and each of the human keypoints on the same side as the target bounding box. From the calculation results, find the two keypoints P1 and P2 closest to the reference point. Record the vertical distance from P1 to P0 as... Simultaneously calculate the vertical distance from P1 to P2. .

[0014] Step 3: Calculate the horizontal distance from P0 to the midline of the human body. And the horizontal distance from P1 to the human body's midline. .

[0015] Step 4: In the standard mannequin diagram, obtain the key points Pm1 and Pm2 corresponding to P1 and P2. Next, calculate the vertical distance from Pm1 to Pm2. And the horizontal distance from Pm1 to the human body's midline .

[0016] Step 5: Calculate the horizontal distance from the center point of the target frame on the puppet diagram to the axis line in the puppet diagram, and the distance from Pm1 on the puppet diagram to the center point of the target frame. The directional distances are denoted as follows: and Following the principle of equal proportions, the proportional relationship between corresponding distances in the puppet diagram and the original diagram is obtained: (1) Calculate using the obtained distance data and Determine the center point of the target box for the non-arm parts in the doll image.

[0017] Step 6: Scale and adjust the width and height of the target frame according to the aspect ratio of the puppet image and the human body frame.

[0018] In one possible implementation, arm part mapping is performed using an arm part mapping strategy, specifically including the following steps: Step 1: After determining whether the target bounding box is located on the left or right side of the human body, select the center point of the target bounding box as the reference point P0'. Next, calculate the vertical distances between the key points on the arm on the same side as the target bounding box and the reference point. Find the two key points P1' and P2' closest to the reference point P0'. Then, obtain the points Pm1' and Pm2' corresponding to these two key points on a standard mannequin drawing. Record the coordinates of each point as P0'. P1' P2' 、Pm1' Pm2' .

[0019] Step 2: Calculate the distances between P1' and P2', P1' and P0', and Pm1' and Pm2' respectively. , and The distance between Pm1' and the center point of the target box on the puppet image is calculated proportionally. .

[0020] Step 3: Calculate the tangent of the angle between the line connecting P1' and P2' and the horizontal direction, the tangent of the angle between the line connecting P1' and P0' and the horizontal direction, and the tangent of the angle between the line connecting Pm1' and Pm2' and the horizontal direction.

[0021] Step 4: According to the principle of similar triangles, the difference between the two included angles in the original image will not change when mapped onto the puppet image. Calculate the angle between the line connecting the center point of the target bounding box and Pm1' on the mapped puppet image and the horizontal direction based on the obtained tangent value. When... (in," When using the XOR operator, the resulting angle needs to be increased by the XOR expression. .

[0022] Step 5: Based on the distance between Pm1' and the center point of the target box mapped onto the puppet diagram, and the angle between the line connecting the center point of the target box mapped onto the puppet diagram and Pm1' and the horizontal direction, calculate the coordinates of the center point of the target box mapped onto the puppet diagram according to the similar triangle relationship.

[0023] Step 6: Scale and adjust the width and height of the target frame according to the aspect ratio of the puppet image and the human body frame.

[0024] Secondly, embodiments of this application provide a millimeter-wave image target mapping device, comprising the following modules: Non-arm part processing module: Adopting a non-arm part mapping strategy, based on the relative positional relationship between corresponding elements in the original image and the puppet image in the vertical and horizontal directions, the center point position of the target box of the non-arm part in the puppet image is calculated, and the target box of the non-arm part is scaled and adjusted.

[0025] Arm part processing module: It adopts an arm part mapping strategy, calculates the center point position of the target box of the arm part in the puppet image based on the relationship between the distance in the original image and the distance between the corresponding points in the puppet image, as well as the principle of similar triangles, and then scales and adjusts the target box of the arm part.

[0026] Mapping module: Based on the center point of the target bounding box in the obtained puppet image, the adjusted non-arm part target bounding box and arm part target bounding box are mapped to the standard puppet image.

[0027] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory; The memory is used to store computer programs.

[0028] When the processor executes the program stored in the memory, it implements any of the millimeter-wave image target mapping methods described in this application.

[0029] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the millimeter-wave image target mapping methods described in this application.

[0030] Fifthly, embodiments of this application provide a computer program product containing instructions that, when run on a computer, cause the computer to execute any of the millimeter-wave image target mapping methods described in this application.

[0031] The beneficial effects of this invention are as follows: To address the different issues that need to be resolved during the mapping process for targets outside the arm and targets outside the arm, different processing strategies are designed for the target bounding boxes outside the arm and targets outside the arm in the mapping method to improve mapping accuracy and visual consistency.

[0032] For target bounding box mapping of non-arm areas, the method no longer relies solely on the proportionality between the straight-line distances between keypoints and the center point of the target bounding box in the original image and the puppet image to determine the position of the center point of the target bounding box in the puppet image. Instead, based on the principle of equal proportions, more precise conditions are used to effectively improve the accuracy of the center point position of the target bounding box in the obtained puppet image.

[0033] For the target bounding box mapping of the arm area, the approach of determining the center point of the target bounding box on the puppet image solely based on distance has been abandoned. Instead, a condition based on the principle of similar triangles has been added. Combined with the condition that these connecting lines are proportional in both the original and puppet images, the accuracy of the determined center point position of the target bounding box on the puppet image can be effectively improved even when the arm is in different poses. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the non-arm part mapping strategy according to an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of the arm part mapping strategy according to an embodiment of the present invention.

[0036] Figure 3 This is a diagram showing the millimeter-wave image target detection results according to an embodiment of the present invention.

[0037] Figure 4 The image shows the result of doll mapping without using the method of this invention.

[0038] Figure 5 This is a diagram showing the result of mannequin mapping using the method of the present invention. Detailed Implementation

[0039] The invention will be further described below with reference to the accompanying drawings.

[0040] The Cartesian coordinate system described below is established with the top-left corner of the image as the origin, where the origin points to the right and downwards respectively. shaft and The positive direction of the axis. It should be noted that the human keypoints and bounding boxes in the millimeter-wave images shown below are obtained using methods such as human pose estimation to describe human pose and structure. These points include, but are not limited to, the head, shoulders, elbows, wrists, torso, hips, knees, and ankles.

[0041] This application provides a millimeter-wave image target mapping method, including the following operations: First, a non-arm part mapping strategy is adopted. Based on the relative positional relationship between corresponding elements in the original image and the puppet image in the vertical and horizontal directions, the center point position of the target box of the non-arm part in the puppet image is calculated, and the target box of the non-arm part is scaled and adjusted.

[0042] The mapping strategy for non-arm areas is as follows: In the original image, the center point of the target bounding box (excluding the arm area), the human body's central axis, and the two closest human body keypoints perpendicularly to the same side of the target bounding box are proportionally related to the center point of the corresponding target bounding box, the human body's central axis, and the two corresponding human body keypoints in the puppet image. Using these existing distances, the center point of the target bounding box (excluding the arm area) in the puppet image is calculated. Then, the target bounding box is scaled and adjusted based on the aspect ratio of the human body bounding box in the puppet image and the original image.

[0043] Secondly, an arm part mapping strategy is adopted. Based on the relationship between the distance in the original image and the distance between the corresponding points in the puppet image, as well as the principle of similar triangles, the center point position of the target box of the arm part in the puppet image is calculated, and the target box of the arm part is scaled and adjusted.

[0044] The arm area mapping strategy is as follows: The distances between the two closest vertical arm keypoints on the same side of the target bounding box in the original image, and the distance between one of these keypoints and the center point of the target bounding box, are proportional to the distances between corresponding points in the puppet image. Furthermore, based on the principle of similar triangles, the difference between the angle between the line connecting the two closest vertical arm keypoints on the same side of the target bounding box in the original image and the horizontal direction, and the angle between the line connecting one of these keypoints and the center point of the target bounding box, remains unchanged after mapping to the puppet image. Using these existing conditions, the center point position of the target bounding box for the arm portion in the puppet image is calculated. Then, the target bounding box is scaled and adjusted based on the aspect ratio of the human body bounding box in the puppet image and the original image.

[0045] Finally, based on the center point of the target bounding box in the obtained puppet image, the adjusted non-arm part target bounding box and arm part target bounding box are mapped onto the standard puppet image.

[0046] In one possible implementation, the non-arm area mapping strategy is as follows, illustrated in the diagram. Figure 1 As shown: Step 1: Calculate the mean of the x-coordinates of the key points of the left and right shoulders and hip joints to obtain the x-coordinate of the human body's midline. Compare the x-coordinate of the human body's midline with the x-coordinate of the target box's center point to determine whether the target box is located on the left or right side of the human body. Alternatively, if the x-coordinates are equal, it is determined that the target box is located on the left side of the human body.

[0047] Step 2: Select the center point P0 of the target bounding box as the reference point, and sequentially traverse the key points on the same side as the target bounding box. Calculate the perpendicularity of the human body key points on the same side as the target bounding box to the reference point P0. ) Directional distance. From the calculation results, find the two key points P1 and P2 that are closest to the reference point P0, and measure the perpendicular distance from P1 to P0. ) Directional distance Simultaneously calculate the perpendicularity from P1 to P2 ( ) Directional distance .

[0048] Step 3: Calculate the horizontal distance from P0 to the midline of the human body. ) Directional distance And the horizontal distance from P1 to the human midline ( ) Directional distance .

[0049] Step 4: In the standard mannequin diagram, obtain the key points Pm1 and Pm2 corresponding to P1 and P2. Then, calculate the vertical distance from Pm1 to Pm2. And the horizontal distance from Pm1 to the human body's midline .

[0050] Step 5: Calculate the horizontal distance from the center point of the target frame on the puppet diagram to the axis line in the puppet diagram, and the distance from Pm1 on the puppet diagram to the center point of the target frame. The directional distances are denoted as follows: and Following the principle of equal proportions, the proportional relationship between corresponding distances in the puppet diagram and the original diagram is obtained: (1) Based on the above known conditions, we can obtain and Then, the coordinates of the center point of the target box mapped onto the puppet diagram according to the proportional relationship are calculated.

[0051] Step 6: Scale and adjust the width and height of the target box according to the ratio of the width and height of the puppet image and the human body frame. This will give you the width and height of the target box in the puppet image.

[0052] In one possible implementation, the arm region mapping strategy is as follows, illustrated in the diagram. Figure 2 As shown: Step 1: Similar to non-arm areas, determine whether the target box is located on the left or right side of the body, then select the center point P0' of the target box as the reference point. Iterate through the key points on the same side as the target box, and calculate the perpendicularity of each key point on the arm (on the same side as the target box) to the reference point P0'. Calculate the directional distance and find the two key locations P1' and P2' closest to the reference point P0'. Then, obtain the corresponding points Pm1' and Pm2' on the standard mannequin diagram, and denote their coordinates as P0'. P1' P2' 、Pm1' Pm2' ; Step 2: Calculate the distances between P1' and P2', P1' and P0', and Pm1' and Pm2' respectively. , and The distance between Pm1' and the center point Pm0' of the target box on the puppet image is calculated proportionally. .

[0053] Step 3: Calculate the angle between the line connecting P1' and P0' and the horizontal direction based on the coordinates of each point in Step 1. tangent value The angle between the line connecting P1' and P2' and the horizontal direction tangent value The angle between the line connecting Pm1' and Pm2' and the horizontal direction tangent value : , , Then, in the original diagram, the angle difference is calculated. ; Step 4: According to the principle of similar triangles, calculate the angle difference before and after mapping. The angle will not change. Based on the tangent value obtained in step 3, the angle between the line connecting the center points Pm0' and Pm1' of the target box mapped onto the puppet diagram and the horizontal direction is calculated. When it appears (in," When using the XOR operator, the resulting angle needs to be increased by the XOR expression. ,Right now .

[0054] Step 5: Based on the distance between Pm1' obtained in Step 2 and the center point Pm0' of the target bounding box mapped onto the puppet image. And the angle between the line connecting the center points Pm0' and Pm1' of the target bounding box obtained in step 4 and the horizontal direction. Calculate the coordinates of point Pm0' on the puppet diagram, which is mapped from the center point of the target bounding box to the point Pm0' based on the similar triangle relationship: , .

[0055] Step 6: Scale and adjust the width and height of the target bounding box according to the aspect ratio of the puppet image and the human body bounding box. This will give you the target bounding box's width and height within the puppet image. At this point, you have obtained the center point position and dimensions of the target bounding box on the puppet image.

[0056] The results of target detection in millimeter-wave images and the results of puppet mapping without using the method of this invention are as follows: Figure 3 As shown in Figure 4, it is clear from the figure that the mapping without the method of this invention resulted in a mapping offset problem for the target box on the left, incorrectly mapping the target box to the arm area. The mapping result of the puppet using the method of this invention is shown in Figure 5. By comparison, it is easy to see that the method of this invention successfully and accurately mapped the target box on the left to the corresponding position on the body.

[0057] For target bounding box mapping of non-arm areas, we no longer rely solely on the proportionality of the straight-line distances between keypoints and the center point of the target bounding box in the original image and the puppet image to determine the position of the target bounding box center point in the puppet image. Instead, based on the principle of equal proportion, we employ a more precise condition: the relative positions of the center point of the target bounding box, the human body's central axis, and the two closest human body keypoints on the same side of the target bounding box, and their corresponding elements in the puppet image, are proportional in both the horizontal and vertical directions. This condition effectively improves the accuracy of the target bounding box center point position in the obtained puppet image.

[0058] For the target bounding box mapping of the arm area, we also abandoned the approach of relying solely on distance to determine the center point of the target bounding box on the puppet image. We added a condition based on the principle of similar triangles: in the original image, we found the two arm keypoints with the closest vertical distance on the same side of the target bounding box, calculated the angle between the line connecting these two keypoints and the horizontal direction, and the angle between the line connecting one of the keypoints and the center point of the target bounding box. The difference between these two angles remains unchanged after mapping to the puppet image. At the same time, combined with the condition that these connecting lines are proportional in the original image and the puppet image, we can effectively improve the accuracy of the determined center point position of the target bounding box on the puppet image even if the arm is in different poses.

[0059] Finally, regardless of whether the target bounding box is for non-arm or arm areas, after mapping the center point position, the mapped target bounding box is scaled and adjusted based on the aspect ratio of the human bounding box in the puppet image and the original image. This operation further improves mapping accuracy and visual consistency.

[0060] This application also provides a millimeter-wave image target mapping device, including the following modules: Non-arm part processing module: Adopting a non-arm part mapping strategy, based on the relative positional relationship between corresponding elements in the original image and the puppet image in the vertical and horizontal directions, the center point position of the target box of the non-arm part in the puppet image is calculated, and the target box of the non-arm part is scaled and adjusted.

[0061] The mapping strategy for non-arm areas is as follows: In the original image, the center point of the target bounding box (excluding the arm area), the human body's central axis, and the two closest human body keypoints perpendicularly to the same side of the target bounding box are proportionally related to the center point of the corresponding target bounding box, the human body's central axis, and the two corresponding human body keypoints in the puppet image. Using these existing distances, the center point of the target bounding box (excluding the arm area) in the puppet image is calculated. Then, the target bounding box is scaled and adjusted based on the aspect ratio of the human body bounding box in the puppet image and the original image.

[0062] Arm part processing module: It adopts an arm part mapping strategy, calculates the center point position of the target box of the arm part in the puppet image based on the relationship between the distance in the original image and the distance between the corresponding points in the puppet image, as well as the principle of similar triangles, and then scales and adjusts the target box of the arm part.

[0063] The arm area mapping strategy is as follows: The distances between the two closest vertical arm keypoints on the same side of the target bounding box in the original image, and the distance between one of these keypoints and the center point of the target bounding box, are proportional to the distances between corresponding points in the puppet image. Furthermore, based on the principle of similar triangles, the difference between the angle between the line connecting the two closest vertical arm keypoints on the same side of the target bounding box in the original image and the horizontal direction, and the angle between the line connecting one of these keypoints and the center point of the target bounding box, remains unchanged after mapping to the puppet image. Using these existing conditions, the center point position of the target bounding box for the arm portion in the puppet image is calculated. Then, the target bounding box is scaled and adjusted based on the aspect ratio of the human body bounding box in the puppet image and the original image.

[0064] Mapping module: Based on the center point of the target bounding box in the obtained puppet image, the adjusted non-arm part target bounding box and arm part target bounding box are mapped to the standard puppet image.

[0065] This application also provides an electronic device, which includes a processor and a memory.

[0066] The memory is used to store computer programs.

[0067] When the processor executes a program stored in the memory, it implements any of the methods described in this application.

[0068] In one possible implementation, the electronic device of this application embodiment further includes a communication interface and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus.

[0069] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc.

[0070] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0071] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0072] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0073] In another embodiment provided in this application, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements any of the methods described in this application.

[0074] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to perform any of the methods described in this application.

[0075] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0076] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0077] The various embodiments in this specification are described in a related manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other.

[0078] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of protection of this application.

Claims

1. A millimeter-wave image target mapping method, characterized in that, Includes the following steps: First, a non-arm part mapping strategy is adopted. Based on the relative positional relationship between corresponding elements in the original image and the puppet image in the vertical and horizontal directions, the center point position of the target box of the non-arm part in the puppet image is calculated, and the target box of the non-arm part is scaled and adjusted. Secondly, an arm part mapping strategy is adopted. Based on the relationship between the distance in the original image and the distance between the corresponding points in the puppet image, as well as the principle of similar triangles, the center point position of the target box of the arm part in the puppet image is calculated, and the target box of the arm part is scaled and adjusted. Finally, based on the center point of the target bounding box in the obtained puppet image, the adjusted non-arm part target bounding box and arm part target bounding box are mapped onto the standard puppet image.

2. A millimeter-wave image target mapping method according to claim 1, characterized in that, The non-arm area mapping strategy is as follows: The relative positions of the center point of the non-arm part of the target box, the human body's central axis, and the two human body key points with the closest vertical distance on the same side of the target box in the original image are proportional to the center point of the target box, the human body's central axis, and the two corresponding human body key points in the puppet image. Using the existing relevant distances, the position of the center point of the non-arm part of the target box in the puppet image is calculated. Then, the target box is scaled and adjusted based on the aspect ratio of the human body box in the puppet image and the original image.

3. A millimeter-wave image target mapping method according to claim 1, characterized in that, The arm part mapping strategy is as follows: The distance between the two closest vertical arm keypoints on the same side of the target box in the original image, and the distance between one of the keypoints and the center point of the target box, are proportional to the distance between the corresponding points in the puppet image. In addition, based on the principle of similar triangles, the difference between the angle between the line connecting the two closest vertical arm keypoints on the same side of the target box and the horizontal direction in the original image, and the angle between the line connecting one of the keypoints and the center point of the target box, remains unchanged after being mapped to the puppet image. Using the existing conditions, the center point of the target box for the arm part in the puppet image is calculated, and then the target box is scaled and adjusted based on the aspect ratio of the puppet image and the human body frame in the original image.

4. A millimeter-wave image target mapping method according to claim 2, characterized in that, Mapping non-arm body parts using a non-arm body part mapping strategy includes the following steps: Step 1: Calculate the x-coordinate of the human body's midline based on the location of key points on the human body, and determine whether the target box is located on the left or right side of the human body; in addition, if the x-coordinates are equal, it is determined that the target box is located on the left side of the human body. Step 2: Select the center point of the target bounding box as the reference point P0; calculate the vertical distance between the reference point and the human body keypoints on the same side as the target bounding box; from the calculation results, find the two keypoints P1 and P2 closest to the reference point; denot the vertical distance from P1 to P0 as... Simultaneously calculate the vertical distance from P1 to P2. ; Step 3: Calculate the horizontal distance from P0 to the midline of the human body. And the horizontal distance from P1 to the human body's midline. ; Step 4: In the standard mannequin diagram, obtain the key points Pm1 and Pm2 corresponding to P1 and P2; then, calculate the vertical distance from Pm1 to Pm2. And the horizontal distance from Pm1 to the human body's midline ; Step 5: Calculate the horizontal distance from the center point of the target frame on the puppet diagram to the axis line in the puppet diagram, and the distance from Pm1 on the puppet diagram to the center point of the target frame. The directional distances are denoted as follows: and Following the principle of equal proportions, the proportional relationship between corresponding distances in the puppet diagram and the original diagram is obtained: (1) Calculate using the obtained distance data and Determine the center point of the target bounding box for the non-arm parts in the doll image; Step 6: Scale and adjust the width and height of the target frame according to the aspect ratio of the puppet image and the human body frame.

5. A millimeter-wave image target mapping method according to claim 3, characterized in that, Arm part mapping is performed using an arm part mapping strategy, which includes the following steps: Step 1: After determining whether the target bounding box is located on the left or right side of the human body, select the center point of the target bounding box as the reference point P0'. Next, calculate the vertical distance between the key points on the arm on the same side as the target bounding box and the reference point. Find the two key points P1' and P2' closest to the reference point P0'. Then, obtain the points Pm1' and Pm2' corresponding to these two key points on a standard mannequin drawing. Record the coordinates of each point as P0'. P1' P2' 、Pm1' Pm2' ; Step 2: Calculate the distances between P1' and P2', P1' and P0', and Pm1' and Pm2' respectively. , and The distance between Pm1' and the center point of the target box on the puppet image is calculated proportionally. ; Step 3: Calculate the tangent of the angle between the line connecting P1' and P2' and the horizontal direction, the tangent of the angle between the line connecting P1' and P0' and the horizontal direction, and the tangent of the angle between the line connecting Pm1' and Pm2' and the horizontal direction, respectively. Step 4: According to the principle of similar triangles, the difference between the two included angles in the original image will not change when mapped onto the puppet image. Calculate the angle between the line connecting the center point of the target bounding box and Pm1' on the puppet image and the horizontal direction based on the obtained tangent value; when... In this case, the obtained included angle needs to be added ,in," " indicates XOR; Step 5: Based on the distance between Pm1' and the center point of the target box mapped onto the puppet diagram, and the angle between the line connecting the center point of the target box mapped onto the puppet diagram and Pm1' and the horizontal direction, calculate the coordinates of the center point of the target box mapped onto the puppet diagram according to the similar triangle relationship. Step 6: Scale and adjust the width and height of the target frame according to the aspect ratio of the puppet image and the human body frame.

6. A millimeter-wave image target mapping device, characterized in that, Includes the following modules: Non-arm part processing module: Adopting a non-arm part mapping strategy, based on the relative positional relationship between corresponding elements in the original image and the puppet image in the vertical and horizontal directions, the center point position of the non-arm part target box in the puppet image is calculated, and the non-arm part target box is scaled and adjusted. Arm part processing module: Adopting an arm part mapping strategy, based on the relationship between the distance in the original image and the distance between the corresponding points in the puppet image, and the principle of similar triangles, the center point position of the target box of the arm part in the puppet image is calculated, and the target box of the arm part is scaled and adjusted. Mapping module: Based on the center point of the target bounding box in the obtained puppet image, the adjusted non-arm part target bounding box and arm part target bounding box are mapped to the standard puppet image.

7. An electronic device, characterized in that, Including processor and memory; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the millimeter-wave image target mapping method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the millimeter-wave image target mapping method according to any one of claims 1-5.

9. A computer program product containing instructions, characterized in that, When it is run on a computer, it causes the computer to perform the millimeter-wave image target mapping method according to any one of claims 1-5.

Images