Millimeter wave image human body anomaly detection method, device, equipment, medium and product

Abstract

The invention discloses a millimeter wave image human body anomaly detection method, device and equipment, a medium and a product. Firstly, human body key points, key point visibility, a human body frame and human body frame confidence in a millimeter wave image are acquired; and secondly, according to the acquired data, whether five human body abnormal conditions of human body area empty, human body blurring, forearm loss, arm non-opening and arm out-of-bound exist in the millimeter wave image is detected, when a certain human body abnormal condition exists in the millimeter wave image, a corresponding error is fed back, and otherwise, the image is fed back to be normal. Related workers can be prompted in time to carry out corresponding processing according to the fed-back error information, invalid processing on images with abnormal human bodies is avoided, therefore, system resources are saved, and the effectiveness and accuracy of subsequent processing are ensured.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of human millimeter wave image target detection, and particularly relates to a millimeter wave image human abnormality detection method, device, equipment, medium and product. BACKGROUND

[0002] Millimeter wave image target detection technology plays a key role in human security inspection field. With excellent contraband detection capability, it has become the preferred security inspection solution in important places such as airports and stations.

[0003] In the process of millimeter wave image target detection, the millimeter wave image is a very critical factor. Once the image is abnormal, it will have a serious negative impact on subsequent target detection. However, in the imaging process of millimeter wave security inspection equipment, improper operation of the examinee often occurs, which often leads to the occurrence of human abnormality in the imaged millimeter wave image, which not only unnecessarily increases the consumption of subsequent system resources, but also easily causes misjudgment and affects the final detection effect.

[0004] In actual application scenarios, the reasons for the occurrence of millimeter wave image human abnormality can be divided into two categories according to whether the examinee shakes during imaging. If the examinee shakes during imaging, it may cause the human body in the millimeter wave image to appear blurred, affecting subsequent target detection. It should be particularly pointed out that the small arm part is most likely to shake, and this part often hides contraband, so special attention should be paid to the situation of this part; if the examinee does not shake during imaging, the examinee's position or posture may be incorrect. Incorrect position, that is, the examinee does not stand in the middle, which may cause the millimeter wave image to basically present a blank background, or the arm exceeds the imaging range, seriously affecting subsequent target detection. As for the posture, since the examinee stands, the shape of the non-arm part is relatively stable and is not easy to cause the image to appear human abnormality, while the arm part may cause abnormality due to the problem of the stretching angle. Specifically, when the examinee's arm stretching angle is too small, strong echo noise interference will be generated between the arm and the side of the body in the imaging result image, affecting subsequent target detection; when the arm stretching angle is too large, the arm will exceed the imaging range, causing the loss of information in this part during detection. SUMMARY

[0005] In view of the problems that the existing millimeter wave image target detection technology is not comprehensive and the accuracy is poor in the image human abnormality detection link, the present application provides a millimeter wave image human abnormality detection method, device, equipment, medium and product, which aims to more comprehensively and accurately judge whether human abnormality occurs in the image.

[0006] In a first aspect, the present application provides a millimeter wave image human abnormality detection method, comprising the following steps:

[0007] The threshold value of the following content and the threshold value The threshold value is obtained by counting a large number of images satisfying "human region empty" , threshold value n and threshold value The threshold value is obtained by counting a large number of images satisfying "human blur" and the threshold value The threshold value is obtained by counting a large number of images satisfying "small arm loss" and the threshold value The threshold value is obtained by counting a large number of images satisfying "arm not extended" and the threshold value The threshold value is obtained by counting a large number of images satisfying "arm out of bounds".

[0008] First, the human key points, key point visibility, human frame and human frame confidence in the millimeter wave image are obtained. The key points include nose, left and right ears, left and right shoulders, left and right elbows, left and right small arms, left and right wrists, left and right hip joints, left and right knees, and left and right ankles.

[0009] Second, according to the obtained data, whether the millimeter wave image exists five kinds of human abnormal conditions of human region empty, human blur, small arm loss, arm not extended and arm out of bounds is detected respectively.

[0010] Finally, when detecting that the millimeter wave image exists a certain human abnormal condition, the corresponding error is fed back, otherwise the image is normal.

[0011] In one possible implementation, the "human region empty" detection operation is as follows:

[0012] Detect whether the human frame confidence in the millimeter wave image is less than the threshold value , and detect whether the gray mean value of the pixels in the human frame region is less than the threshold value . If any of the two conditions is met, it is determined that the millimeter wave image exists "human region empty".

[0013] In one possible implementation, the "human blur" detection operation is as follows:

[0014] Detect whether the number of key points with visibility value less than the threshold value in the millimeter wave image, except for nose, left and right ears and left and right ankles, exceeds the threshold value n; and detect whether the standard deviation of the pixels in the body region surrounded by left and right shoulders and left and right hip joints is less than the threshold value . If any of the above conditions is met, it is determined that the millimeter wave image exists "human blur".

[0015] In one possible implementation, the "forearm missing" detection operation is as follows:

[0016] Detect the visibility value of each side's forearm and wrist key point in the millimeter wave image. If the visibility value of any one of the points is lower than a threshold value or the sum of the visibility values of the two points is lower than a threshold value and at the same time the forearm region is free of prohibited objects, it is determined that the millimeter wave image has "forearm missing".

[0017] In one possible implementation, the "arm not extended" detection operation is as follows:

[0018] Calculate the angle of the upper arm with respect to the vertical direction according to the two key points of each side's shoulder and elbow in the millimeter wave image, and calculate the angle of the forearm with respect to the vertical direction according to the two key points of each side's elbow and wrist. Detect whether the angle of the upper arm with respect to the vertical direction is lower than a threshold value and whether the angle of the forearm with respect to the vertical direction is lower than a threshold value . If any one side satisfies any one of the two conditions, it is determined that the millimeter wave image has "arm not extended".

[0019] In one possible implementation, the "arm out of bounds" detection operation is as follows:

[0020] Detect the visibility value of each side's wrist key point in the millimeter wave image. If the visibility value of any one of the wrist key points is found to be lower than a threshold value and the distance between the side human body frame boundary and the image boundary is not more than a threshold value , it is determined that the millimeter wave image has "arm out of bounds".

[0021] In a second aspect, the embodiments of the present application provide a millimeter wave image human body anomaly detection device, comprising the following modules:

[0022] Data acquisition module: used for acquiring human body key points, key point visibility, human body frame and human body frame confidence in the millimeter wave image. The key points include nose, left and right ears, left and right shoulders, left and right elbows, left and right forearms, left and right wrists, left and right hip joints, left and right knees, and left and right ankles.

[0023] "Human body region empty" detection module: based on the data acquired by the data acquisition module, detect whether the human body frame confidence in the millimeter wave image is less than a threshold value and whether the gray mean value of the human body frame region pixels is less than a threshold value . If any one of the two conditions is satisfied, it is determined that the millimeter wave image has "human body region empty".

[0024] "Human body blur" detection module: based on the data obtained by the data acquisition module, detect whether the number of key points in the millimeter wave image, other than the nose, left and right ears, and left and right ankles, whose visibility value is lower than the threshold value exceeds the threshold value n; at the same time, detect whether the standard deviation of the pixels in the body region surrounded by the left and right shoulders and the left and right hip joints is less than the threshold value . If any of the above conditions is met, it is determined that the millimeter wave image has "human body blur".

[0025] "Small arm loss" detection module: based on the data obtained by the data acquisition module, detect the visibility value of each side small arm and wrist key point in the millimeter wave image, if the visibility value of any one point is lower than the threshold value or the sum of the visibility values of the two points is lower than the threshold value , and at the same time, the small arm region does not contain prohibited objects, it is determined that the millimeter wave image has "small arm loss".

[0026] "Arm not extended" detection module: based on the data obtained by the data acquisition module, calculate the angle of the large arm with the vertical direction according to the two key points of each side shoulder and elbow in the millimeter wave image, and calculate the angle of the small arm with the vertical direction according to the two key points of each side elbow and wrist, detect whether the angle of the large arm with the vertical direction is lower than the threshold value , and whether the angle of the small arm with the vertical direction is lower than the threshold value . If any of the two conditions is met on either side, it is determined that the millimeter wave image has "arm not extended".

[0027] "Arm out of bounds" detection module: based on the data obtained by the data acquisition module, detect the visibility value of each side wrist key point in the millimeter wave image, if the visibility value of any one side wrist key point is found to be lower than the threshold value , and the distance between the boundary of the body frame of that side and the image boundary does not exceed the threshold value , it is determined that the millimeter wave image has "arm out of bounds".

[0028] Threshold determination module: used for statistical calculation to determine the threshold values, by statistically calculating a large number of images that meet "human body region empty" to obtain threshold values and threshold value , by statistically calculating a large number of images that meet "human body blur" to obtain threshold values , threshold value n and threshold value , by statistically calculating a large number of images that meet "small arm loss" to obtain threshold values and threshold value , by statistically calculating a large number of images that meet "arm not extended" to obtain threshold values and threshold value , by statistically calculating a large number of images that meet "arm out of bounds" to obtain threshold values and threshold .

[0029] The result output module is used to output corresponding error information when any detection module detects a human abnormality in the millimeter-wave image. If no human abnormality is detected in the millimeter-wave image, the module outputs the message "Image normal".

[0030] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory;

[0031] The memory is used to store computer programs.

[0032] When the processor executes the program stored in the memory, it implements any of the millimeter-wave image human anomaly detection methods described in this application.

[0033] 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 human anomaly detection methods described in this application.

[0034] 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 human anomaly detection methods described in this application.

[0035] The beneficial effects of this invention are as follows:

[0036] To address the shortcomings of existing millimeter-wave image target detection technologies in detecting human anomalies, such as incomplete detection and low accuracy, this invention provides a millimeter-wave image human anomaly detection method. This method aims to more comprehensively and accurately determine whether human anomalies exist in images. Based on whether the person being inspected moves during the imaging process, the causes of human anomalies are categorized into two types. These types are further subdivided into five specific scenarios: empty human body area, blurred human body, missing forearm, arm not extended, and arm outside the image boundary. This approach also allows for timely prompting of relevant personnel to handle errors based on feedback, avoiding ineffective processing of images with human anomalies, thus saving system resources and ensuring the effectiveness and accuracy of subsequent processing. Attached Figure Description

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other embodiments can also be obtained by those skilled in the art based on these drawings.

[0038] Figure 1 Flow chart of the method embodiment of the present application.

[0039] Figure 2 Related legend of the millimeter wave image human body information.

[0040] Figure 3 Normal human body millimeter wave image detected by the present application.

[0041] Figure 4 Abnormal human body millimeter wave image detected by the present application. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art based on the present application belong to the scope of protection of the present application.

[0043] The method embodiment of the present application divides the reasons for causing human body abnormality into two categories according to whether the examinee shakes during the imaging process, and further divides the human body abnormality into five specific situations on the basis of the two categories. The core idea of the method is to respectively judge whether the image exists five human body abnormal situations, i.e. human body region empty, human body blur, small arm loss, arm not unfolded and arm out of boundary. If it is detected that the image exists a certain human body abnormality, the corresponding error is fed back. If it is detected that the image does not exist human body abnormality, the image is normal is fed back.

[0044] The method embodiment of the present application provides a millimeter wave image human body abnormality detection method, as shown in Figure 1 , including the following steps:

[0045] Firstly, the human body key points, key point visibility, human body frame and human body frame confidence in the millimeter wave image are acquired. The key points include nose, left and right ears, left and right shoulders, left and right elbow, left and right small arm, left and right wrist, left and right hip joint, left and right knee and left and right ankle. The related example image is as shown in Figure 1 .

[0046] Secondly, according to the acquired data, it is detected whether the millimeter wave image exists five human body abnormal conditions including human body region empty, human body blur, small arm loss, arm not unfolded and arm out of boundary.

[0047] Finally, when it is detected that the millimeter wave image exists certain human body abnormality, the corresponding error is fed back, otherwise the image is normal.

[0048] In a possible implementation, the "human body region empty" detection operation is as follows:

[0049] It is detected whether the human body frame confidence is less than =0.8, and whether the average gray value of the human body frame region pixels is less than =30. If any of the two conditions is met, it is determined that the millimeter wave image exists "human body region empty", and error code 1 is returned.

[0050] In a possible implementation, the "human body blur" detection operation is as follows:

[0051] It is detected whether the number of key points other than the nose, left and right ears and left and right ankles in the millimeter wave image, whose visibility value is less than =0.8, exceeds n=4, and whether the standard deviation of the pixels in the body region surrounded by the left and right shoulders and left and right hip joints is less than =18. If any of the above conditions is met, it is determined that the millimeter wave image exists "human body blur", and error code 2 is returned.

[0052] In a possible implementation, the "small arm loss" detection operation is as follows:

[0053] The visibility value of each side small arm and wrist key point in the millimeter wave image is detected, if the visibility value of any one point on either side is less than =0.4 or the sum of the visibility values of the two points is less than =1.2, and at the same time the small arm region does not exist prohibited objects, it is determined that the millimeter wave image exists "small arm loss", and error code 3 is returned.

[0054] In a possible implementation, the "arm not unfolded" detection operation is as follows:

[0055] According to the two key points of each side shoulder and elbow in the millimeter wave image, the angle between the large arm and the vertical direction is calculated, and according to the two key points of each side elbow and wrist, the angle between the small arm and the vertical direction is calculated, it is detected whether the angle between the large arm and the vertical direction is less than =15°, and whether the angle between the small arm and the vertical direction is less than threshold = If either of the two conditions is met for any side, it is determined that the millimeter wave image has "arms not unfolded", and error code 4 is returned.

[0056] In a possible implementation, the "arms out of bounds" detection operation is as follows:

[0057] The visibility value of each side wrist key point in the millimeter wave image is detected, and if the visibility value of any side wrist key point is found to be lower than = 0.1, and the distance between the side human body frame boundary and the image boundary is not more than = 20 pixels, it is determined that the millimeter wave image has "arms out of bounds", and error code 5 is returned.

[0058] The embodiments of the present application also provide a millimeter wave image human body anomaly detection device, comprising the following modules:

[0059] A data acquisition module is configured to acquire human body key points, key point visibility, a human body frame, and human body frame confidence in a millimeter wave image. The key points include a nose, left and right ears, left and right shoulders, left and right elbows, left and right forearms, left and right wrists, left and right hip joints, left and right knees, and left and right ankles.

[0060] A "human body region empty" detection module is configured to detect, based on the data acquired by the data acquisition module, whether the human body frame confidence in the millimeter wave image is less than a threshold value = 0.8, and whether the average gray value of the pixels in the human body frame region is less than a threshold value = 30. If any of the two conditions is met, it is determined that the millimeter wave image has "human body region empty".

[0061] A "human body blur" detection module is configured to detect, based on the data acquired by the data acquisition module, whether the number of key points, other than the nose, left and right ears, and left and right ankles, in the millimeter wave image, whose visibility value is lower than a threshold value = 0.8, exceeds a threshold value n = 4; and whether the standard deviation of the pixels in the body region surrounded by the left and right shoulders and the left and right hip joints is less than a threshold value = 18. If any of the above conditions is met, it is determined that the millimeter wave image has "human body blur".

[0062] A "small arm loss" detection module is configured to detect, based on the data acquired by the data acquisition module, the visibility value of each side small arm and wrist key point in the millimeter wave image, and if any side has one key point whose visibility value is lower than a threshold value = 0.4 or the sum of the visibility values of the two key points is lower than a threshold value = 1.2, and at the same time, the small arm region does not have prohibited objects, it is determined that the millimeter wave image has "small arm loss".

[0063] The "Arm Not Open" detection module: Based on data acquired by the data acquisition module, it calculates the angle of the upper arm relative to the vertical direction using two key points (shoulder and elbow) in the millimeter-wave image on each side, and calculates the angle of the forearm relative to the vertical direction using two key points (elbow and wrist) on each side. It then detects whether the angle of the upper arm relative to the vertical direction is below a threshold. =15°, is the angle between the forearm and the vertical lower than the threshold? =0°. If either side satisfies either of the two conditions, then the millimeter-wave image is determined to have an "arm not open".

[0064] The "arm out of bounds" detection module: Based on data acquired by the data acquisition module, it detects the visibility value of key points on each wrist in the millimeter-wave image. If the visibility value of any key point on either wrist is found to be below a threshold, the module will detect the out-of-bounds wrist. =0.1, and the distance between the human body bounding box boundary and the image boundary on this side does not exceed the threshold. If the value is 20, then the millimeter-wave image is determined to have an "arm out of bounds".

[0065] Threshold determination module: Used for statistical calculation to determine the threshold, which is obtained by statistically analyzing a large number of images that meet the "empty human body region" condition. =0.8 and threshold =30, the threshold was obtained by statistically analyzing a large number of images that satisfy the "human body blur" condition. =0.8, threshold n=4 and threshold =18, the threshold was obtained by statistically analyzing a large number of images that meet the "forearm loss" criterion. =0.4 and threshold =1.2, the threshold was obtained by statistically analyzing a large number of images that meet the condition of "arms not open". =15° and threshold =0°, the threshold is obtained by statistically analyzing a large number of images that meet the "arm out of bounds" condition. =0.1 and threshold =20.

[0066] The result output module is used to output corresponding error information when any detection module detects a human abnormality in the millimeter-wave image. If no human abnormality is detected in the millimeter-wave image, the module outputs the message "Image normal".

[0067] The normal and abnormal human millimeter-wave images detected according to the embodiments of this application are as follows: Figure 3 and Figure 4 As shown in the figure, it is clear from the image that when a millimeter-wave image of a normal human body is detected, the device returns code 0; when a millimeter-wave image of an abnormal human body is detected, the device returns the corresponding error code (…). Figure 4The returned error code is 2, indicating that the human body is blurred, thereby prompting the relevant staff to process.

[0068] The electronic device provided by the embodiment of the present application further includes a processor and a memory.

[0069] The memory is used for storing a computer program.

[0070] The processor is used for executing the program stored on the memory, and realizes any method provided in the present application.

[0071] In a possible implementation, the electronic device provided by the embodiment of the present application further includes a communication interface and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the communication bus.

[0072] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0073] The communication interface is used for communication between the above electronic device and other devices.

[0074] The memory can include a Random Access Memory (RAM) and can also include a Non-Volatile Memory (NVM), for example, at least one disk memory. Optionally, the memory can also be at least one storage device located away from the aforementioned processor.

[0075] The processor mentioned above can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0076] In a further implementation provided in the present application, a computer readable storage medium is also provided, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method described in any of the present application.

[0077] In a further implementation provided in the present application, a computer program product containing instructions, which when run on a computer, causes the computer to perform the method described in any of the present application.

[0078] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented 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 the present application are generated. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (for example, floppy disk, hard disk, magnetic tape), optical media (for example, DVD), or semiconductor media (for example, solid state disk (SSD)) and the like.

[0079] It should be noted that in this paper, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitation, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0080] Each of the embodiments in the specification is described in a related manner, and each embodiment focuses on the difference from other embodiments. The same and similar parts between the embodiments can be referred to each other.

[0081] The above only describes the preferred embodiments of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for millimeter wave image human anomaly detection, characterized in that, The method comprises the following steps: First, acquire the human body key points, key point visibility, human body frame and human body frame confidence in the millimeter wave image; the key points include nose, left and right ears, left and right shoulders, left and right elbow, left and right forearm, left and right wrist, left and right hip joint, left and right knee and left and right ankle; Second, detect whether the millimeter wave image exists five human body abnormal conditions of human body region empty, human body blur, forearm loss, arm not unfolded and arm out of bounds according to the acquired data; Finally, when it is detected that the millimeter wave image exists a certain human body abnormal condition, feedback the corresponding error, otherwise feedback the image is normal. 2.The method of claim 1, wherein, The "human body region empty" detection operation is as follows: Detecting whether a human body frame confidence in a millimeter wave image is less than a threshold , and detecting whether a gray mean value of a human body frame region pixel is less than a threshold ; if any one of the two conditions is met, it is determined that the millimeter wave image has a "human body region empty". 3.The method of claim 1, wherein, The "human body blur" detection operation is as follows: Visibility values of other key points in the millimeter wave image, except for nose, left and right ears, and left and right ankles, are less than a threshold value whether the number of key points exceeds a threshold value n; and whether the standard deviation of pixels in a body region surrounded by left and right shoulders and left and right hip joints is less than a threshold value If any of the above conditions is met, it is determined that the millimeter wave image has "human body blur". 4.The method of claim 1, wherein, The "forearm loss" detection operation is as follows: Detecting the visibility value of each side of the forearm and wrist key points in the millimeter wave image, if any one of the points on either side has a visibility value below a threshold value Or the sum of the visibility values of the two points is below a threshold value And at the same time, the forearm area does not exist prohibited, it is determined that the millimeter wave image exists "forearm loss".

5. The method of claim 1, wherein, The "arm not unfolded" detection operation is as follows: According to the two key points of each side shoulder and elbow in the millimeter wave image, the angle of the upper arm with the vertical direction is calculated, and according to the two key points of each side elbow and wrist, the angle of the lower arm with the vertical direction is calculated, and whether the angle of the upper arm with the vertical direction is lower than the threshold value is detected , whether the angle of the lower arm with the vertical direction is lower than the threshold value ; if any side satisfies any one of the two conditions, it is determined that the millimeter wave image exists "arm not open".

6. The method of claim 1, wherein, The "arm out of bounds" detection operation is as follows: detecting a visibility value of each side wrist key point in the millimeter wave image, if finding that the visibility value of any side wrist key point is lower than a threshold value , and the distance between the side human body frame boundary and the image boundary does not exceed a threshold value , determining that the millimeter wave image exists "arm out of bounds".

7. A millimeter wave image human anomaly detection apparatus, characterized by, The method comprises the following modules: The data acquisition module is used for acquiring the human body key points, key point visibility, human body frame and human body frame confidence in the millimeter wave image; the key points include nose, left and right ears, left and right shoulders, left and right elbow, left and right forearm, left and right wrist, left and right hip joint, left and right knee and left and right ankle; The "human region empty" detection module detects whether the confidence of the human box in the millimeter wave image is less than a threshold based on the data obtained by the data acquisition module , and simultaneously detects whether the average gray value of the pixels in the human box region is less than a threshold . If any one of the two conditions is met, it is determined that the millimeter wave image exists "human body region empty"; "Human Blur" Detection Module: Based on data acquired by the data acquisition module, this module detects key points in millimeter-wave images, excluding the nose, left and right ears, and left and right ankles, where the visibility value is below a threshold. Check if the number of key points exceeds the threshold n; simultaneously check if the standard deviation of pixels in the body region enclosed by the left and right shoulders and left and right hip joints is less than the threshold. If any of the above conditions are met, the millimeter-wave image is determined to have "human body blur". "Missing arm" detection module: based on the data obtained by the data acquisition module, the visibility value of each side of the arm and wrist key point in the millimeter wave image is detected, and if any side has a visibility value of one point below the threshold Or the sum of the visibility values of the two points is below the threshold , and at the same time, the arm area does not exist prohibited, it is determined that the millimeter wave image exists "missing arm"; "Arm not extended" detection module: based on the data obtained by the data acquisition module, calculating the angle between the upper arm and the vertical direction according to the two key points of each side shoulder and elbow in the millimeter wave image, and calculating the angle between the lower arm and the vertical direction according to the two key points of each side elbow and wrist, detecting whether the angle between the upper arm and the vertical direction is lower than the threshold value , whether the angle between the lower arm and the vertical direction is lower than the threshold value ; if any one side satisfies any one of the two conditions, it is determined that the millimeter wave image exists "arm not extended"; An "arm out of bound" detection module: based on the data acquired by the data acquisition module, detect the visibility value of each side wrist key point in the millimeter wave image, if it is found that the visibility value of any side wrist key point is lower than the threshold , and the distance between the side human body frame boundary and the image boundary does not exceed the threshold , it is determined that the millimeter wave image exists "arm out of bound"; Threshold determination module: used for statistical calculation to determine threshold, by counting a large number of images satisfying "human region empty" to obtain threshold and threshold , by counting a large number of images satisfying "human blur" to obtain threshold , threshold n and threshold , by counting a large number of images satisfying "small arm loss" to obtain threshold and threshold , by counting a large number of images satisfying "arm not extended" to obtain threshold and threshold , by counting a large number of images satisfying "arm out of frame" to obtain threshold and threshold ; The result output module is used for outputting the error information of the corresponding human body abnormality when any detection module detects that the millimeter wave image exists human body abnormality; if it is detected that the millimeter wave image does not exist any human body abnormal condition, output the information of "image is normal".

8. An electronic device, comprising: The processor and the memory are included; The memory is used for storing computer programs; The processor is used for executing the programs stored in the memory, and realizes the millimeter wave image human body abnormality detection method in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer program is stored in the computer readable storage medium, and the computer program is executed by the processor to realize the millimeter wave image human body abnormality detection method in any one of claims 1-6.

10. A computer program product comprising instructions, characterized in that, When it runs on the computer, it makes the computer execute the millimeter wave image human body abnormality detection method in any one of claims 1-6.

Images