Camera mounting quality analysis method and device

Abstract

The embodiment of the application provides a camera erection quality analysis method and device, and relates to the technical field of video monitoring.The method comprises the following steps: determining a target statistical mode set for a target camera; wherein the target statistical mode is a preset mode for analyzing multiple faces; determining a specified image photographed by the target camera in the target statistical mode; acquiring a deflection angle of a face in the specified image with respect to a specified direction; and generating an analysis result for the erection quality of the target camera based on the acquired deflection angle.Through the scheme, the camera erection quality analysis can be more accurate and efficient.

Description

Technical Field

[0001] This invention relates to the field of video surveillance technology, and in particular to a method and apparatus for analyzing the quality of camera installation. Background Technology

[0002] Nowadays, cameras are widely deployed in various locations, such as places with high or low foot traffic. Furthermore, after installation, the angle of the cameras may deviate, resulting in tilted images of the captured scenes.

[0003] To optimize camera installation, it is usually necessary to analyze the installation quality and adjust the camera angle based on the analysis results.

[0004] It is evident that how to accurately and efficiently analyze the quality of camera installation is an urgent problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide a method and apparatus for analyzing the quality of camera installation, so as to achieve accurate and efficient analysis of camera installation quality. The specific technical solution is as follows:

[0006] In a first aspect, embodiments of the present invention provide a method for analyzing the quality of camera installation, comprising:

[0007] Determine the target statistical mode set for the target camera; wherein the target statistical mode is a preset mode for analyzing multiple faces;

[0008] Acquire the specified image captured by the target camera under the target statistical mode;

[0009] Obtain the deflection angle of a face in the specified image with respect to a specified direction;

[0010] Based on the obtained deflection angle, an analysis result is generated regarding the installation quality of the target camera.

[0011] Optionally, acquiring the specified image captured by the target camera under the target statistical mode includes:

[0012] Determine the image selection parameters corresponding to the target statistical pattern; wherein, different statistical patterns correspond to different image selection parameters;

[0013] Based on the image selection parameters, a specified image captured by the target camera is selected.

[0014] Optionally, the target statistical mode is a multi-frame mode or a single-frame mode;

[0015] The image selection parameters corresponding to the multi-frame mode include: the number of image frames and / or the acquisition time range;

[0016] The image selection parameters corresponding to the single-frame mode include: a threshold for the number of faces present in the image.

[0017] Optionally, the target statistical mode is a fixed number of faces mode or a fixed duration mode;

[0018] The image selection parameters corresponding to the fixed number of faces mode include: the number of target faces;

[0019] The image selection parameters corresponding to the fixed duration mode include: acquisition duration.

[0020] Optionally, selecting the specified image captured by the target camera based on the image selection parameters includes:

[0021] If the target mode is a multi-frame mode, multiple frames that meet the specified frame number and / or acquisition time range are selected from the images captured by the target camera to obtain multiple specified images.

[0022] Optionally, before selecting multiple frames from the images captured by the target camera that conform to the specified frame number and / or acquisition time range to obtain multiple specified frames, the method further includes:

[0023] Determine the auxiliary parameters corresponding to the multi-frame mode; wherein, the auxiliary parameters include: a threshold for the number of faces present in the image;

[0024] From the images captured by the target camera, multiple frames that meet the specified frame number and / or acquisition time range are selected to obtain multiple specified images, including:

[0025] From the images captured by the target camera, select multiple frames that meet the specified frame count and / or acquisition time range, and where the number of faces contained in each image is not less than the specified threshold, to obtain multiple specified images.

[0026] Optionally, selecting the specified image captured by the target camera based on the image selection parameters includes:

[0027] If the target statistical mode is a single-frame mode, a frame image containing no less than the number threshold of faces is selected from the images captured by the target camera to obtain a specified frame image.

[0028] Optionally, selecting the specified image captured by the target camera based on the image selection parameters includes:

[0029] If the target statistical mode is a fixed number of faces mode, at least one frame image is selected from the images captured by the target camera, the total number of faces being equal to the target number of faces, to obtain the specified image.

[0030] Optionally, selecting the specified image captured by the target camera based on the image selection parameters includes:

[0031] If the target statistical mode is a fixed duration mode, at least one frame of image captured by the target camera within the acquisition duration is selected from the images captured by the target camera to obtain at least one specified frame of image.

[0032] Optionally, obtaining the deflection angle of the face in the specified image about the specified direction includes:

[0033] The omnidirectional face detection algorithm is used to detect the deflection angle of a face in the specified image with respect to a specified direction.

[0034] Optionally, the deflection angle of the face about a specified direction includes: the deflection angle of the face about the vertical direction of the image, the deflection angle of the face about the vertical line direction, or the deflection angle of the face about the vertical direction of the image display screen.

[0035] Optionally, generating analysis results regarding the installation quality of the target camera based on the acquired deflection angle includes:

[0036] Based on the obtained deflection angle, the deviation angle of the target camera during installation is determined.

[0037] Optionally, determining the deviation angle of the target camera during installation based on the acquired deflection angle includes:

[0038] The obtained deflection angle is processed by a specified calculation to obtain the deviation angle of the target camera during installation; wherein the specified calculation includes averaging or mode calculation.

[0039] Optionally, the method further includes:

[0040] When the deviation angle exceeds a predetermined threshold, the adjustment direction and / or adjustment angle corresponding to the deviation angle are determined, and a camera adjustment notification containing the adjustment direction and / or adjustment angle is output.

[0041] or,

[0042] When the deviation angle exceeds a predetermined threshold, a camera adjustment notification containing the deviation angle is output.

[0043] Secondly, embodiments of the present invention also provide a camera installation quality analysis device, the device comprising:

[0044] The first determining module is used to determine the target statistical mode set for the target camera; wherein the target statistical mode is a preset mode for analyzing multiple faces;

[0045] The second determining module is used to acquire a specified image captured by the target camera under the target statistical mode;

[0046] The acquisition module is used to acquire the deflection angle of a face in the specified image about a specified direction;

[0047] The generation module is used to generate analysis results on the installation quality of the target camera based on the acquired deflection angle.

[0048] Thirdly, embodiments of the present invention also provide an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0049] Memory, used to store computer programs;

[0050] When the processor executes the program stored in the memory, it implements the camera installation quality analysis method provided in the first aspect above.

[0051] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the camera installation quality analysis method provided in the first aspect above.

[0052] This invention also provides a computer program product containing instructions that, when run on a computer, causes the computer to execute any of the camera installation quality analysis methods described above.

[0053] Beneficial effects of the embodiments of the present invention:

[0054] This invention provides a method for analyzing the installation quality of a camera. First, a target statistical mode is determined for the target camera. Then, a specified image captured by the target camera under the target statistical mode is determined. Next, the deflection angle of faces in the specified image about a specified direction is obtained. Based on the obtained deflection angle, an analysis result for the installation quality of the target camera is generated. It is evident that this solution, during quality analysis, uses the deflection angles of multiple faces in a specified direction, avoiding errors caused by relying on a single face, thus ensuring accurate analysis of installation quality. Furthermore, different statistical modes are set for monitoring scenarios with different pedestrian traffic, allowing different modes to be set for target cameras in different scenarios, thereby enabling adaptive analysis for different scenarios. Compared to using a uniform analysis method for different scenarios, this solution, while ensuring multi-face analysis, is more closely aligned with the characteristics of the scenario itself, undoubtedly making the installation quality analysis more efficient. Therefore, this solution allows for accurate and efficient analysis of camera installation quality.

[0055] Of course, implementing any product or method of the present invention does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0057] Figure 1 This is a flowchart of the camera installation quality analysis method provided in an embodiment of the present invention;

[0058] Figure 2 This is a schematic diagram of the face deflection angle in an embodiment of the present invention;

[0059] Figure 3 This is a schematic diagram of a rectangular frame representing the position of a face in an embodiment of the present invention;

[0060] Figure 4 This is another flowchart of the camera installation quality analysis method provided in the embodiments of the present invention;

[0061] Figure 5 This is another flowchart of the camera installation quality analysis method provided in the embodiments of the present invention;

[0062] Figure 6 A schematic diagram of the camera mounting quality analysis device provided in an embodiment of the invention;

[0063] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0064] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of the present invention.

[0065] Nowadays, cameras are widely deployed in various public places. However, after some cameras are installed, the angle may be deviated or even wrong, and the selection of installation points may be unreasonable. These installation problems are difficult to detect and require a certain amount of manpower.

[0066] To accurately and efficiently analyze the quality of camera installation, embodiments of the present invention provide a method and apparatus for analyzing camera installation quality. The method for analyzing camera installation quality provided in these embodiments will be described first.

[0067] The camera installation quality analysis method provided in this invention can be applied to electronic devices. In specific applications, the electronic device can be a server or a terminal device, both of which are reasonable. In practical applications, the terminal device can be a tablet computer, desktop computer, etc.

[0068] Furthermore, the entity executing this camera installation quality analysis method can be a camera installation quality analysis device. For example, when the camera installation quality analysis method is applied to a terminal device, the camera installation quality analysis device can be functional software running on the terminal device for analyzing and processing camera installation quality; when the camera installation quality analysis method is applied to a server, the camera installation quality analysis device can be a computer program running on the server.

[0069] An embodiment of the present invention provides a method for analyzing the installation quality of a camera, which may include the following steps:

[0070] Determine the target statistical mode set for the target camera; wherein the target statistical mode is a preset mode for analyzing multiple faces;

[0071] Determine the specified image captured by the target camera under the target statistical mode;

[0072] Obtain the deflection angle of a face in the specified image with respect to a specified direction;

[0073] Based on the obtained deflection angle, an analysis result is generated regarding the installation quality of the target camera.

[0074] The camera installation quality analysis method provided in this invention first determines the target statistical mode set for the target camera; then, it acquires a specified image captured by the target camera under the target statistical mode; next, it acquires the deflection angle of the face in the specified image with respect to a specified direction; and based on the acquired deflection angle, it generates an analysis result for the installation quality of the target camera. It is evident that in this solution, during quality analysis, the deflection angle of multiple faces in a specified direction avoids errors caused by relying on a single face, ensuring accurate analysis of installation quality. Furthermore, different statistical modes are set for monitoring scenarios with different pedestrian traffic, allowing different modes to be set for target cameras in different scenarios, thus enabling adaptive analysis for different scenarios. Compared to using a uniform analysis method for different scenarios, this solution, while ensuring multi-face analysis, is more closely aligned with the characteristics of the scenario itself, undoubtedly making the installation quality analysis more efficient. Therefore, this solution allows for accurate and efficient analysis of camera installation quality.

[0075] The following describes a method for analyzing the quality of camera installation provided by an embodiment of the present invention, with reference to the accompanying drawings.

[0076] like Figure 1 As shown, the camera installation quality analysis method provided in this embodiment of the invention may include steps S101-S104:

[0077] S101, determine the target statistics mode set for the target camera; wherein, the target statistics mode is a preset mode for analyzing multiple faces;

[0078] The target camera can be any camera required for installation quality analysis, such as those used in security monitoring, video conferencing, or telemedicine. Furthermore, it is understood that since this solution relies on facial analysis for camera installation quality analysis, the target camera is typically located in a scene where faces can be captured. Moreover, the collection, storage, use, and processing of facial data comply with relevant laws and regulations and do not violate public order and good morals.

[0079] In this embodiment, to avoid analysis errors caused by a single face, a multi-face analysis method is used to ensure the accuracy of the quality analysis. Furthermore, the number of faces that can be contained in a single frame varies depending on the scene and the volume of people. For example, in scenes with high traffic volume, a single frame can contain multiple faces simultaneously, eliminating the need to analyze multiple frames. Conversely, in scenes with low traffic volume, multiple frames are required to contain multiple faces. Therefore, to perform efficient quality analysis, this embodiment pre-sets multiple statistical modes for multi-face analysis. These different statistical modes are designed for monitoring different traffic volumes. This allows for the selection of different statistical modes based on different scenes, avoiding the inefficiency of using a fixed image analysis method regardless of the scene, and thus enabling efficient and adaptive analysis for different scenarios.

[0080] Based on the above processing approach, for the target camera to be subjected to quality analysis, before conducting the quality analysis, a statistical mode can be set for the target camera based on the scene in which the target camera is located. In this way, during the quality analysis process, the target statistical mode set for the target camera can be determined first, and then subsequent processing can be performed based on the target statistical mode. There are multiple ways to set the statistical mode for any camera, and this embodiment of the invention does not limit this. For example, the statistical mode can be set for the camera through the human-computer interaction interface provided by the electronic device, thereby allowing the electronic device to directly determine the statistical mode set for the camera; or, the scene type in which the camera is located can be set through the human-computer interaction interface provided by the electronic device, thereby allowing the electronic device to determine the statistical mode set for the camera based on a pre-defined mapping relationship between scene types and statistical modes.

[0081] For example, in one implementation, the multiple statistical modes may include a multi-frame mode and a single-frame mode. In this case, the target statistical mode can be either a single-frame mode or a multi-frame mode selected based on the actual scene. For instance, in a scene with low pedestrian traffic, the target statistical mode can be set to a multi-frame mode, that is, collecting multiple frames of images for face analysis to obtain a sufficient number of samples and improve accuracy; in a scene with high pedestrian traffic, the target statistical mode can be set to a single-frame mode, that is, using a single frame of image captured by the camera for face analysis.

[0082] For example, in another implementation, multiple statistical modes may include: a fixed number of faces mode and a fixed duration mode. In the fixed number of faces mode, images are continuously acquired until the number of identified faces reaches a specified threshold; all images acquired during this period are used as the designated images for quality analysis. This mode can be applied to scenarios with low pedestrian traffic, thus obtaining a sufficient number of samples by setting a fixed number of faces, avoiding the problem of insufficient faces for quality analysis in low-traffic situations. In the fixed duration mode, images captured by the target camera within a specified duration are used as the designated images; this statistical mode can be applied to scenarios with high pedestrian traffic, as a sufficient number of samples can be obtained within the specified duration.

[0083] It should be noted that the application scenarios of the above statistical models are merely illustrative and do not constitute a limitation on the application scenarios. Furthermore, when analyzing multiple faces, the specific representation forms of the patterns included in the multiple statistical models can vary, and the embodiments of this invention do not impose any limitations.

[0084] S102, acquire the specified image captured by the target camera under the target statistical mode;

[0085] That is, based on the target statistical mode and the actual scene, a specified image is selected from the images captured by the target camera for use in setting up quality analysis. The number of specified images required may vary depending on the statistical mode. After determining the target statistical mode, at least one specified image captured by the target camera can be determined for use in setting up quality analysis.

[0086] Optionally, in one implementation, acquiring the specified image captured by the target camera under the target statistical mode may include steps A1-A2:

[0087] Step A1: Determine the image selection parameters corresponding to the target statistical pattern; wherein, different statistical patterns correspond to different image selection parameters;

[0088] Different image selection parameters are pre-set for different statistical modes to adapt to different scenarios.

[0089] Step A2: Based on the image selection parameters, select the specified image captured by the target camera.

[0090] Understandably, the number of images used for quality analysis varies depending on the statistical model, and therefore the number of specified images selected based on the image selection parameters also varies.

[0091] For example, when the target statistics mode is multi-frame mode or single-frame mode;

[0092] The image selection parameters corresponding to the multi-frame mode may include: the number of image frames and / or the acquisition time range;

[0093] That is, select a specified number of frames or select multiple frames within the acquisition time range for use in quality analysis;

[0094] The image selection parameters corresponding to the single-frame mode may include: a threshold for the number of faces present in the image.

[0095] That is, a frame of image with a number of faces reaching a threshold is selected for quality analysis.

[0096] When the target statistical mode is a fixed number of faces mode or a fixed duration mode;

[0097] The image selection parameters corresponding to the fixed number of faces mode may include: the number of target faces;

[0098] That is, the specified images to be selected are determined based on the required number of target faces.

[0099] The image selection parameters corresponding to the fixed duration mode may include: acquisition duration.

[0100] That is, the images captured by the target camera within the acquisition time are selected as the designated images.

[0101] The following text will provide examples of how to select a specific image for the four statistical modes mentioned above, and will not be repeated here.

[0102] S103, Obtain the deflection angle of the face in the specified image about the specified direction;

[0103] Ideally, if the target camera is set up without any angular deviation, the central axis of the face should be exactly located in the vertical direction of the image, the vertical line direction, or the vertical direction of the image display screen. Therefore, when the face faces the camera, the angle of deflection of the face's central axis relative to the vertical direction of the image, the vertical line direction, or the vertical direction of the image display screen, i.e., the roll angle, can reflect the deviation angle of the target camera. Therefore, in this embodiment, the deflection angle of the face about a specified direction can include: the deflection angle of the face about the vertical direction of the image, the deflection angle of the face about the vertical line direction, or the deflection angle of the face about the vertical direction of the image display screen. For example, as shown... Figure 2 As shown, Figure 2The angle of deflection of the face's central axis relative to the vertically upward direction in the image is called clockwise deflection theta. In practical applications, a clockwise deflection angle can be considered positive, and a counter-clockwise deflection angle can be considered negative, and vice versa. It should be noted that the method of representing the deflection angle includes, but is not limited to, this; any method that can represent the face's roll angle is acceptable. Furthermore, the determination of the deflection angle of the face in the specified image with respect to the specified direction can be automatically obtained by the camera setup quality analysis device. Of course, in a monitoring scenario, other monitoring tasks may exist. If other monitoring tasks have already analyzed the deflection angle of the face in each frame with respect to the specified direction, then the camera setup quality analysis device can obtain the deflection angle of the face in the specified image with respect to the specified direction from the analysis results of the other monitoring tasks, thus making the setup quality analysis more efficient. Of course, the process of selecting the specified image can also rely on the processing results of other monitoring tasks, which is also reasonable.

[0104] Understandably, typical face detection determines the position and size of a face in an image, using a 4D array to represent the face's position and size as a bounding box. For example, a 4D array (x, y, w, h) can be used to represent the face's position and size, where x and y represent the coordinates of the center point of the bounding box, w represents the width of the bounding box, and h represents the height of the bounding box. In this embodiment, to obtain the deflection angle in a specified direction, one or more dimensions can be added to the face detection results. These added dimensions can then be used to characterize the deflection angle in the specified direction. Thus, by performing face detection on a given image, the deflection angle in the specified direction can be determined from the face detection results.

[0105] For example, in one implementation, the detection results of face detection can be represented using a five-dimensional array. This array can represent not only the location and size of the face, but also the deflection angle with respect to a specified direction. For instance, the five-dimensional array could be (x, y, w, h, theta), where x and y represent the coordinates of the rectangle, w represents the width of the rectangle, h represents the height of the rectangle, and theta represents the deflection angle with respect to a specified direction; or (x1, y1, x2, y2, theta), where x1 and y1 represent the coordinates of the top-left corner of the rectangle, x2 and y2 represent the coordinates of the bottom-right corner, and theta represents the deflection angle with respect to a specified direction.

[0106] For example, in another implementation, the face detection result can be represented using a six-dimensional array. For instance, the face detection result can be represented by an offset of an inscribed rectangle plus an upright rectangle. In this case, the face detection result can be a six-dimensional array (x, y, w, h, Δα, Δβ), such as... Figure 3 As shown, x and y represent the coordinates of the center point of the circumscribed rectangle, w represents the width of the circumscribed rectangle, h represents the height of the circumscribed rectangle, and Δα and Δβ represent the offsets of the inscribed rectangle relative to the center point of the circumscribed rectangle. At this point, the roll angle of the face can be calculated using the formula... The conclusion is as follows.

[0107] Furthermore, in practical applications, the target camera may be installed upside down, resulting in inverted faces in the captured image. Therefore, to ensure the detection of both horizontal and inverted faces, and to address various degrees of camera installation deviation, an omnidirectional face detection algorithm can be used to detect the angle of rotation of faces in a specified image relative to a specified direction. This omnidirectional face detection algorithm originates from remote sensing image detection technology. While detecting an object, it outputs the object's angle of rotation relative to the image, representing the object's position, size, and angle in 5-8 dimensions. In this embodiment, when using the omnidirectional face detection algorithm, a five-dimensional representation can be used. For example, an angle parameter theta can be appended to the position coordinates (x, y, w, h), representing the clockwise rotation angle relative to the y-axis of the Cartesian coordinate system, i.e., the angle of rotation in the specified direction. It is understandable that omnidirectional face detection algorithms can solve the following problems: Face detection based on MTCNN (Multi-task convolutional neural network) cannot detect faces when the face orientation is horizontal, inverted, etc., and the range of angles that can be processed is limited.

[0108] It should be noted that if the target camera is a camera in a face recognition scenario and the above-mentioned omnidirectional face detection algorithm is used for face detection, then the acquisition of the deflection angle in the camera installation quality analysis method provided in this embodiment of the invention is a link in the entire camera system. Therefore, the installation quality analysis of the target camera has almost no additional overhead and will not bring significant computational costs.

[0109] Furthermore, any method capable of obtaining the deflection angle of a face in the specified image with respect to a specified direction can be applied to this embodiment.

[0110] S104, Based on the acquired deflection angle, generate analysis results on the installation quality of the target camera.

[0111] The analysis results can serve as a reference for adjusting the installation angle and location of the target camera. This solution can automatically generate analysis results on the installation quality of the target camera without requiring extensive manual measurement and screening of the camera's deflection angle, thus effectively saving labor costs.

[0112] The analysis results generated based on the acquired deflection angle regarding the installation quality of the target camera may include:

[0113] Based on the obtained deflection angle, the deviation angle of the target camera during installation is determined.

[0114] At this point, a quality analysis of the target camera's setup is performed, including determining the target camera's deviation angle. It is understood that the acquired face deflection angle can be used as a reference for adjusting the target camera's setup angle. For example, if the face's roll angle in the image is clockwise theta, then the target camera's deviation angle can be determined to be counterclockwise theta.

[0115] There are multiple ways to determine the deviation angle of the target camera during installation based on the obtained deflection angle.

[0116] In one implementation, determining the deviation angle of the target camera during installation based on the acquired deflection angle may include:

[0117] The obtained deflection angle is processed by a specified calculation to obtain the deviation angle of the target camera during installation; wherein the specified calculation includes averaging or mode calculation.

[0118] When performing specified calculations, clockwise deflection angles can be considered positive, and counterclockwise deflection angles can be considered negative, facilitating calculations; the reverse is also possible. By averaging or calculating the mode of the deflection angles, errors caused by a single deflection angle can be avoided, improving the accuracy of the analysis of the target camera's installation quality.

[0119] For a given calculation, the average value can be directly calculated. For example, if the number of images is 20 frames, and 10 faces are detected in frames 1 to 10, and 15 faces are detected in frames 11 to 20, then the deflection angles of all faces in these 20 frames are counted. A total of 10*10 + 10*15 = 250 deflection angles are summed and then divided by the total number of faces to obtain the average deflection angle of the faces in these 20 frames. This average deflection angle is then used as the deviation angle of the target camera during setup.

[0120] The specified operation can also be a mode operation. For example, for the multiple deflection angles obtained, they can be divided into 5-degree increments. If the number of faces is most in the 10-15 degree range, then the deflection angle of the target camera is considered to be 10-15 degrees.

[0121] There are many common calculation and processing methods, which will not be listed one by one in this invention. By performing specified calculation and processing on each obtained deflection angle, statistical methods can be used to eliminate individual errors, avoid deviations in face deflection angles caused by individual postures, improve the accuracy of setup quality analysis, and not be affected by significant changes in face quality over time (such as night and day). It has strong robustness to errors caused by individual differences and environmental changes over time.

[0122] The camera installation quality analysis method provided by the present invention may further include:

[0123] When the deviation angle exceeds a predetermined threshold, the adjustment direction and / or adjustment angle corresponding to the deviation angle are determined, and a camera adjustment notification containing the adjustment direction and / or adjustment angle is output.

[0124] or,

[0125] When the deviation angle exceeds a predetermined threshold, a camera adjustment notification containing the deviation angle is output.

[0126] After performing specified calculations on each deflection angle, the installation deviation angle of the target camera can be obtained. It is understood that a small installation deviation angle is not worth adjusting. Therefore, a threshold angle, such as 10 degrees, can be set for the installation deviation angle. When the installation angle is less than the threshold, it can be considered a statistical error, and even if it is a true deviation, it will not affect use. When the threshold is exceeded, i.e., the deviation angle exceeds the predetermined threshold, a camera adjustment notification containing the deviation angle can be output to warn maintenance personnel that the installation angle of the target camera needs calibration. The output method of the camera adjustment notification includes, but is not limited to, text or voice. Calibration methods include, but are not limited to, manual adjustment, target camera motor calibration, etc. Furthermore, the adjustment direction and / or adjustment angle corresponding to the deviation angle can be determined based on the deviation angle, and a camera adjustment notification containing the adjustment direction and / or adjustment angle can be output. For example, if the deflection angle is 15 degrees counterclockwise, then the corresponding adjustment direction is clockwise, the adjustment angle is 15 degrees, and a camera adjustment notification containing the adjustment direction and / or adjustment angle is output again.

[0127] The camera installation quality analysis method provided in this invention first determines the target statistical mode set for the target camera; then, it determines a specified image captured by the target camera under the target statistical mode; next, it obtains the deflection angle of the face in the specified image with respect to a specified direction; and based on the obtained deflection angle, it generates an analysis result for the installation quality of the target camera. It is evident that in this solution, during quality analysis, the deflection angle of multiple faces in a specified direction avoids errors caused by relying on a single face, ensuring accurate analysis of installation quality. Furthermore, different statistical modes are set for monitoring scenarios with different pedestrian traffic, allowing different modes to be set for target cameras in different scenarios, thus enabling adaptive analysis for different scenarios. Compared to using a uniform analysis method for different scenarios, this solution, while ensuring multi-face analysis, is more closely aligned with the characteristics of the scenario itself, undoubtedly making the installation quality analysis more efficient. Therefore, this solution allows for accurate and efficient analysis of camera installation quality.

[0128] The following describes the specific implementation of selecting a specified image captured by the target camera based on the image selection parameters, using the target statistical mode as a multi-frame mode:

[0129] Optionally, in one implementation, the target statistical mode is a multi-frame mode; the image selection parameters corresponding to the multi-frame mode include: the number of image frames and / or the acquisition time range;

[0130] At this time, selecting a specified image captured by the target camera based on the image selection parameters may include:

[0131] From the images captured by the target camera, select multiple frames that meet the specified frame number and / or acquisition time range to obtain multiple specified images.

[0132] Regarding the image selection parameter being the number of image frames, when selecting a specified image, multiple frames captured by the target camera can be selected; alternatively, a frame-sampling method can be used to select multiple frames to obtain the specified images. For example, multiple specified images can be selected from a video captured by the target camera over the past twenty minutes, by extracting one frame every 23 frames. The specific value of the number of image frames is not specifically limited in this embodiment of the invention, provided that multiple faces can be obtained.

[0133] When the image selection parameter is the acquisition time range, multiple frames of images captured within the predetermined acquisition time range can be selected to obtain multiple specified images. For example, the acquisition time range can also be a specified time period, such as one hour from now, or 8:00-9:00 every day.

[0134] Furthermore, for image selection parameters based on the acquisition time range, the aforementioned specified calculation processing can be weighted averaging. In one implementation, the weights can decay over time; in another, the weights can be set to higher values ​​during peak traffic periods and lower values ​​at other times. For example, if the acquisition time range is 8:00 to 12:00, with a weight of 1.0 during peak traffic periods (8:00 to 9:00) and 0.1 during off-peak periods (9:00 to 12:00), and if 1000 faces with a deviation angle of 1 degree appear between 8:00 and 9:00, while only 10 faces with a deviation angle of -10 degrees appear between 9:00 and 12:00, then the deviation angle obtained after weighted averaging is: (1*1000*1+0.1*10*(-10)) / (1*1000+0.1*10)=0.899. Further analysis revealed that the deviation angle of the target camera was 0.899 degrees. It can be seen that in this case, the deviation angle is small and not worth adjusting, so no warning is required.

[0135] In addition, regarding the aforementioned multi-frame mode, in one implementation, before selecting multiple frames from the images captured by the target camera that conform to the specified number of image frames and / or acquisition time range to obtain multiple specified images, the following may be included:

[0136] Determine the auxiliary parameters corresponding to the multi-frame mode; wherein, the auxiliary parameters include: a threshold for the number of faces present in the image;

[0137] At this time, selecting multiple frames from the images captured by the target camera that meet the specified frame number and / or acquisition time range to obtain multiple specified images includes:

[0138] From the images captured by the target camera, select multiple frames that meet the specified frame count and / or acquisition time range, and where the number of faces contained in each image is not less than the specified threshold, to obtain multiple specified images.

[0139] In this implementation, an auxiliary parameter corresponding to the multi-frame mode is added to ensure that the number of faces in each specified image frame reaches the threshold. For example, when the number of image frames is set to 200, the acquisition time range is from the current time, and the auxiliary parameter is 2, the number of faces in the current image is determined starting from the first image captured by the target camera. If the number of faces in the current image is less than 2, the next image frame is judged. If the number of faces in the current image reaches 2, it is selected as the specified image. The above process is repeated until the number of selected specified images reaches 200.

[0140] In this embodiment, for the target statistical mode being a multi-frame mode, the image selection parameters set include: the number of image frames and / or the acquisition time range. From the images captured by the target camera, multiple frames that meet the specified number of image frames and / or acquisition time range are selected to obtain multiple specified images. Through this scheme, it can be ensured that a sufficient number of face images are acquired in the multi-frame mode for setup quality analysis, thereby improving the accuracy of setup quality analysis.

[0141] The following describes the specific implementation of selecting a specified image captured by the target camera based on the image selection parameters, using the target statistics mode as a single-frame mode:

[0142] Optionally, in one implementation, the target statistical mode is a single-frame mode; the image selection parameters may include a threshold for the number of faces present in the image;

[0143] At this time, selecting the specified image captured by the target camera based on the image selection parameters may include:

[0144] From the images captured by the target camera, select a frame of a specified image containing at least the number of faces.

[0145] To improve the accuracy of quality analysis, in scenarios with high pedestrian traffic, only one frame with a number of faces not less than a certain threshold can be selected from the images captured by the target camera as the designated image. This threshold can be set according to specific circumstances. In single-frame mode, a start time can also be set; for example, starting from the current moment, only one frame with a number of faces not less than the threshold can be selected from the images captured by the target camera as the designated image.

[0146] In this embodiment, for the target statistical mode being single-frame mode, the image selection parameters include: the image selection parameters include a threshold for the number of faces present in the image, thereby selecting a specified frame image from the image captured by the target camera whose number of faces is not less than the threshold; through this scheme, it can be ensured that a sufficient number of face images are collected in single-frame mode for setup quality analysis, thereby improving the accuracy of setup quality analysis.

[0147] The following describes the specific implementation of selecting a specified image captured by the target camera based on the image selection parameters, using a fixed number of faces as the target statistical mode:

[0148] Optionally, in one implementation, the target statistical mode includes a fixed number of faces mode; the image selection parameters may include: the target number of faces;

[0149] The step of selecting a specified image captured by the target camera based on the image selection parameters includes:

[0150] From the images captured by the target camera, at least one frame is selected whose total number of faces equals the target number of faces to obtain a specified image.

[0151] In this statistical mode, images are continuously selected based on the set target number of faces, regardless of the size of the crowd, until the number of faces in the specified images reaches the target number. For example, if the target number of faces is set to 1000, images captured by the target camera are continuously selected, and the number of faces in each selected frame is detected until the total number of faces in the selected images reaches 1000.

[0152] In this embodiment, for the target statistical mode of fixed face count, at least one frame of image is selected from the images captured by the target camera, where the total number of faces is equal to the target number of faces, to obtain a specified image. This method can obtain a sufficient number of face images for installation quality analysis, regardless of the size of the crowd, thereby improving the accuracy of installation quality analysis.

[0153] The following describes the specific implementation method of selecting a specified image captured by the target camera based on the image selection parameters, using the target statistical mode as a fixed duration mode:

[0154] Optionally, in one implementation, the target statistical mode includes a fixed number of faces mode; the image selection parameters may include: acquisition duration;

[0155] The step of selecting a specified image captured by the target camera based on the image selection parameters includes:

[0156] From the images captured by the target camera, at least one frame of the image captured by the target camera within the acquisition time period is selected to obtain at least one specified image.

[0157] In this statistical mode, the start time for data collection needs to be preset, for example, set within a time range with high pedestrian traffic. All images captured within the specified collection period can be used as the designated images, or a frame-by-frame extraction method can be used to select designated images, for example, extracting one frame every 23 frames. Furthermore, the collection period can be set according to the pedestrian traffic volume. When pedestrian traffic is high, the collection period can be appropriately reduced; when pedestrian traffic is low, the collection period can be appropriately increased.

[0158] In this embodiment, for the target statistical mode being a fixed duration mode, at least one frame of image captured by the target camera within the specified acquisition duration is selected from the images captured by the target camera to obtain at least one specified image. It can be seen that in this mode, by adjusting the acquisition duration, a sufficient number of face images can be obtained for setup quality analysis, thereby improving the accuracy of the setup quality analysis.

[0159] To facilitate understanding of the camera installation quality analysis method provided in the embodiments of the present invention, an exemplary description is given below in conjunction with the accompanying drawings.

[0160] For example, such as Figure 4 As shown, the target statistical mode is pre-set to multi-frame mode, and the acquisition time range and the threshold for the number of faces in the auxiliary parameters are set. After data collection begins, faces in each frame of the image acquired by the target camera are continuously detected to obtain face detection results. The detection results include the number of faces detected in each frame and the deflection angle of each face. It is determined whether the number of faces in each frame meets the limit of the auxiliary parameters. If not, the next frame is detected. If yes, the calculated deflection angle in that image is included in the statistical pool. At the same time, it is continuously determined whether the acquisition time of the target camera meets the statistical duration set in the acquisition time range. If not, the next frame is acquired. If yes, the statistical mean of the deflection angles in the statistical pool is calculated to obtain the deviation angle of the target camera.

[0161] Based on the above processing approach, the processing flow of the camera installation quality analysis method provided in the embodiments of the present invention in practical applications can be as follows: Figure 5 As shown:

[0162] First, the video data collected by the target camera is input into the omnidirectional face detection module. The omnidirectional face detection module is used to determine the specified image based on the target statistical pattern and image selection parameters, and to obtain the detection box information of the face in the specified image using the omnidirectional face detection algorithm, that is, the coordinates and size of the rectangle that encloses the face, and the face angle information, that is, the face deflection angle. Then, the face angle information and the detection box information are input into the statistics module.

[0163] The statistics module uses the input deflection angle and the number of deflection angles to perform specified calculations to obtain the deviation angle of the target camera during installation and generate adjustment suggestions for the target camera. For example, if the deviation angle of the target camera during installation is clockwise theta, the adjustment suggestion could be to adjust the target camera counterclockwise theta. For target cameras used for face detection, if the number of faces recognized over a long period of time is too small, the adjustment suggestion could also be to reselect the installation location.

[0164] Finally, the collected detection frame information, face angle information, and adjustment suggestions can be sent to the backend for manual review.

[0165] As can be seen, this solution, during quality analysis, avoids errors caused by relying on a single face by using the deflection angle of multiple faces in a specified direction, thus ensuring accurate analysis of installation quality. Furthermore, different statistical modes are set for monitoring scenarios with varying pedestrian traffic, allowing for different modes to be applied to target cameras in different scenarios. This enables adaptive analysis for different scenarios. Compared to using a uniform analysis method for different scenarios, this solution, while ensuring multi-face analysis, is more closely aligned with the characteristics of the specific scenario, undoubtedly making installation quality analysis more efficient. Therefore, this solution allows for accurate and efficient analysis of camera installation quality.

[0166] Corresponding to the above method embodiments, this invention also provides a camera installation quality analysis device, such as... Figure 6 As shown, the device includes:

[0167] The first determining module 610 is used to determine the target statistical mode set for the target camera; wherein the target statistical mode is a preset mode for analyzing multiple faces;

[0168] The second determining module 620 is used to acquire a specified image captured by the target camera under the target statistical mode;

[0169] The acquisition module 630 is used to acquire the deflection angle of a face in the specified image about a specified direction;

[0170] The generation module 640 is used to generate analysis results on the installation quality of the target camera based on the acquired deflection angle.

[0171] Optionally, the second determining module 620 includes:

[0172] The first determining submodule is used to determine the image selection parameters corresponding to the target statistical pattern; wherein, different statistical patterns correspond to different image selection parameters;

[0173] The selection submodule is used to select a specified image captured by the target camera based on the image selection parameters.

[0174] Optionally, the target statistical mode is a multi-frame mode or a single-frame mode;

[0175] The image selection parameters corresponding to the multi-frame mode include: the number of image frames and / or the acquisition time range;

[0176] The image selection parameters corresponding to the single-frame mode include: a threshold for the number of faces present in the image.

[0177] Optionally, the target statistical mode is a fixed number of faces mode or a fixed duration mode;

[0178] The image selection parameters corresponding to the fixed number of faces mode include: the number of target faces;

[0179] The image selection parameters corresponding to the fixed duration mode include: acquisition duration.

[0180] Optionally, the target statistics mode is a multi-frame mode;

[0181] The selection submodule is specifically used for:

[0182] If the target mode is a multi-frame mode, multiple frames that meet the specified frame number and / or acquisition time range are selected from the images captured by the target camera to obtain multiple specified images.

[0183] Optionally, the second determining module further includes:

[0184] The second determining submodule is used to determine auxiliary parameters corresponding to the multi-frame mode before selecting multiple frames that meet the specified frame number and / or acquisition time range from the images captured by the target camera by the selection submodule to obtain multiple specified frames; wherein, the auxiliary parameters include: a threshold for the number of faces present in the image;

[0185] The selection submodule is specifically used for:

[0186] From the images captured by the target camera, select multiple frames that meet the specified frame count and / or acquisition time range, and where the number of faces contained in each image is not less than the specified threshold, to obtain multiple specified images.

[0187] Optionally, the target statistics mode is a single-frame mode;

[0188] The selection submodule is specifically used for:

[0189] If the target statistical mode is a single-frame mode, a frame image containing no less than the number threshold of faces is selected from the images captured by the target camera to obtain a specified frame image.

[0190] Optionally, the target statistical pattern includes a fixed number of faces pattern;

[0191] The selection submodule is specifically used for:

[0192] If the target statistical mode is a fixed number of faces mode, at least one frame image is selected from the images captured by the target camera, the total number of faces being equal to the target number of faces, to obtain the specified image.

[0193] Optionally, the target statistical pattern includes a fixed duration pattern;

[0194] The selection submodule is specifically used for:

[0195] If the target statistical mode is a fixed duration mode, at least one frame of image captured by the target camera within the acquisition duration is selected from the images captured by the target camera to obtain at least one specified frame of image.

[0196] Optionally, the acquisition module is specifically used for:

[0197] The omnidirectional face detection algorithm is used to detect the deflection angle of a face in the specified image with respect to a specified direction.

[0198] Optionally, the deflection angle of the face about a specified direction includes: the deflection angle of the face about the vertical direction of the image, the deflection angle of the face about the vertical line direction, or the deflection angle of the face about the vertical direction of the image display screen.

[0199] Optionally, the generation module is specifically used for:

[0200] Based on the obtained deflection angle, the deviation angle of the target camera during installation is determined.

[0201] Optionally, the generation module determines the deviation angle of the target camera during installation based on the acquired deflection angle, including:

[0202] The obtained deflection angle is processed by a specified calculation to obtain the deviation angle of the target camera during installation; wherein the specified calculation includes averaging or mode calculation.

[0203] Optionally, the device further includes:

[0204] The output module is used to determine the adjustment direction and / or adjustment angle corresponding to the deviation angle when the deviation angle exceeds a predetermined threshold, and output a camera adjustment notification containing the adjustment direction and / or adjustment angle.

[0205] or,

[0206] When the deviation angle exceeds a predetermined threshold, a camera adjustment notification containing the deviation angle is output.

[0207] This invention also provides an electronic device, such as... Figure 7 As shown, it includes a processor 701, a communication interface 702, a memory 703, and a communication bus 704, wherein the processor 701, the communication interface 702, and the memory 703 communicate with each other through the communication bus 704.

[0208] Memory 703 is used to store computer programs;

[0209] When the processor 701 executes the program stored in the memory 703, it implements the steps of the above-mentioned camera installation quality analysis method.

[0210] 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. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0211] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0212] 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.

[0213] 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.

[0214] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor, implements the steps of the above-described camera installation quality analysis method.

[0215] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute the camera installation quality analysis method described in the above embodiment.

[0216] 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 the present invention 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)).

[0217] 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.

[0218] The various embodiments in this specification are described in a related manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0219] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for analyzing the quality of camera installation, characterized in that, The method includes: A target statistical mode is determined for the target camera; wherein the target statistical mode is a preset mode for analyzing multiple faces; and for the scene in which the target camera is located, the image acquired under the target statistical mode for analysis contains multiple faces. Acquire a specified image captured by the target camera under the target statistical mode; wherein, when the target statistical mode is a multi-frame mode, the specified image is a multi-frame image captured by the target camera within the acquisition time range; the acquisition time range is the image selection parameter corresponding to the multi-frame mode; An omnidirectional face detection algorithm is used to detect the deflection angle of a face in a specified image about a specified direction; wherein, the deflection angle of the face about the specified direction is the roll angle; the detection result of the face detection obtained by the omnidirectional face detection algorithm is represented by a six-dimensional array, wherein the six-dimensional array includes the coordinates of the center point of an upright rectangle, the width of the upright rectangle, the height of the upright rectangle, and the offset of the face detection box relative to the center point of the upright rectangle; the upright rectangle is the bounding rectangle of the face detection box; the deflection angle of the face in the specified image about the specified direction is calculated based on the width of the upright rectangle, the height of the upright rectangle, and the offset of the face detection box relative to the center point of the upright rectangle; The obtained deflection angles are weighted and averaged to obtain the deviation angle of the target camera during installation; wherein, the weight of the deflection angle obtained based on the specified image collected during peak traffic hours is higher than the weight of the deflection angle obtained based on the specified image collected at other times.

2. The method according to claim 1, characterized in that, The step of acquiring the specified image captured by the target camera under the target statistical mode includes: Determine the image selection parameters corresponding to the target statistical pattern; wherein, different statistical patterns correspond to different image selection parameters; Based on the image selection parameters, a specified image captured by the target camera is selected.

3. The method according to claim 2, characterized in that, The target statistical mode is either a multi-frame mode or a single-frame mode; The image selection parameters corresponding to the multi-frame mode also include: the number of image frames; The image selection parameters corresponding to the single-frame mode include: a threshold for the number of faces present in the image.

4. The method according to claim 2, characterized in that, The target statistical mode is either a fixed number of faces mode or a fixed duration mode; The image selection parameters corresponding to the fixed number of faces mode include: the number of target faces; The image selection parameters corresponding to the fixed duration mode include: acquisition duration.

5. The method according to claim 3, characterized in that, The step of selecting a specified image captured by the target camera based on the image selection parameters includes: If the target statistical mode is a multi-frame mode, multiple frames that meet the specified number of frames and the acquisition time range are selected from the images captured by the target camera to obtain multiple specified images.

6. The method according to claim 5, characterized in that, Before selecting multiple frames of images that meet the specified frame number and acquisition time range from the images captured by the target camera to obtain multiple specified frames, the method further includes: Determine the auxiliary parameters corresponding to the multi-frame mode; wherein, the auxiliary parameters include: a threshold for the number of faces present in the image; The process of selecting multiple frames from the images captured by the target camera that meet the specified frame number and acquisition time range to obtain multiple specified images includes: From the images captured by the target camera, select multiple frames that meet the specified frame number and acquisition time range, and where the number of faces contained in each image is not less than the specified threshold, to obtain multiple specified images.

7. The method according to claim 3, characterized in that, The step of selecting a specified image captured by the target camera based on the image selection parameters includes: If the target statistical mode is a single-frame mode, a frame image containing no less than the number threshold of faces is selected from the images captured by the target camera to obtain a specified frame image.

8. The method according to claim 4, characterized in that, The step of selecting a specified image captured by the target camera based on the image selection parameters includes: If the target statistical mode is a fixed number of faces mode, at least one frame image is selected from the images captured by the target camera, the total number of faces being equal to the target number of faces, to obtain the specified image.

9. The method according to claim 4, characterized in that, The step of selecting a specified image captured by the target camera based on the image selection parameters includes: If the target statistical mode is a fixed duration mode, at least one frame of image captured by the target camera within the acquisition duration is selected from the images captured by the target camera to obtain at least one specified frame of image.

10. The method according to any one of claims 1-8, characterized in that, The deflection angle of the face with respect to a specified direction includes: the deflection angle of the face with respect to the vertical direction of the image, the deflection angle of the face with respect to the vertical direction of the vertical line, or the deflection angle of the face with respect to the vertical direction of the image display screen.

11. The method according to claim 1, characterized in that, The method further includes: When the deviation angle exceeds a predetermined threshold, the adjustment direction and / or adjustment angle corresponding to the deviation angle are determined, and a camera adjustment notification containing the adjustment direction and / or adjustment angle is output. or, When the deviation angle exceeds a predetermined threshold, a camera adjustment notification containing the deviation angle is output.

12. A camera installation quality analysis device, characterized in that, The device includes: The first determining module is used to determine the target statistical mode set for the target camera; wherein, the target statistical mode is a preset mode for analyzing multiple faces; and for the scene where the target camera is located, the image acquired for analysis under the target statistical mode contains multiple faces. The second determining module is used to acquire a specified image captured by the target camera under the target statistical mode; wherein, when the target statistical mode is a multi-frame mode, the specified image is a multi-frame image captured by the target camera within the acquisition time range; the acquisition time range is the image selection parameter corresponding to the multi-frame mode; The acquisition module is used to detect the deflection angle of a face in a specified image about a specified direction using an omnidirectional face detection algorithm; wherein, the deflection angle of the face about the specified direction is the roll angle; the detection result of the face detection obtained by the omnidirectional face detection algorithm is represented by a six-dimensional array, the six-dimensional array including the coordinates of the center point of an upright rectangle, the width of the upright rectangle, the height of the upright rectangle, and the offset of the face detection box relative to the center point of the upright rectangle; the upright rectangle is the bounding rectangle of the face detection box; the deflection angle of the face in the specified image about the specified direction is calculated based on the width of the upright rectangle, the height of the upright rectangle, and the offset of the face detection box relative to the center point of the upright rectangle; The generation module is used to perform weighted average processing on the acquired deflection angles to obtain the deviation angle of the target camera during installation; wherein, the weight of the deflection angle obtained based on the specified image collected during peak traffic hours is higher than the weight of the deflection angle obtained based on the specified image collected at other times.

Images