WO2020094091A1

WO2020094091A1 - Image capturing method, monitoring camera, and monitoring system

Info

Publication number: WO2020094091A1
Application number: PCT/CN2019/116219
Authority: WO
Inventors: 王晶晶
Original assignee: 杭州海康威视数字技术股份有限公司
Priority date: 2018-11-07
Filing date: 2019-11-07
Publication date: 2020-05-14
Also published as: CN111163259A

Abstract

Provided in the embodiments of the present application are an image capturing method, a monitoring camera, and a monitoring system. The image capturing method comprises: collecting a current video frame; performing human facial target detection on the current video frame to determine a human facial target image of each human facial target in the current video frame; and aiming at a designated human facial target, taking the human facial target image with optimal image quality within a preset capturing interval as a captured image of the designated human facial target when the current video frame satisfies the preset capturing interval, and uploading same to a comparison system. According to the present solution, it can be ensured that a comparison result of the comparison system has higher accuracy.

Description

Image capturing method, monitoring camera and monitoring system

This application requires the priority of the Chinese patent application filed on November 7, 2018 in the Chinese Patent Office, with the application number 201811321515.6 and the invention titled "An Image Capture Method, Surveillance Camera and Surveillance System" In this application.

Technical field

The present application relates to the technical field of video surveillance, in particular to an image capture method, surveillance camera and surveillance system.

Background technique

In order to reduce the social crime rate, a large number of surveillance cameras are installed in all corners of the city. The surveillance camera captures the face object appearing in the scene, and uploads the captured image to the comparison system. The comparison system extracts the face features from the captured image, and extracts the extracted face features and the black list face features Perform a comparison, and if the similarity of the comparison is greater than a certain threshold, an alarm is generated.

Whenever the surveillance camera detects a face target, it will capture the face target. However, the face target generally appears continuously in the video, and the surveillance camera will capture the captured image in multiple consecutive video frames. All are uploaded to the comparison system, which brings huge pressure to the transmission and storage of the captured images and the data processing of the comparison system.

In order to deal with the above problems, in the corresponding image capture method, after the surveillance camera detects the face target, the face target is captured every certain number of video frames, then the comparison system can receive the captured image at a fixed frequency , Greatly reducing the number of captured images uploaded by the surveillance camera to the comparison system. However, in each captured image uploaded by the surveillance camera, the face target may be poorly recognizable due to blurring, occlusion, and other reasons. If the recognizable face target in each captured image received by the system is If it is poor, it will affect the similarity of the comparison, causing false alarms or false negatives, resulting in lower accuracy of the comparison results.

Summary of the invention

The purpose of the embodiments of the present application is to provide an image capture method, a monitoring camera, and a monitoring system, so as to ensure that the comparison result of the comparison system has high accuracy. The specific technical solutions are as follows:

In a first aspect, an embodiment of the present application provides an image capture method, which includes:

Collect the current video frame;

Perform face target detection on the current video frame to determine the face target image of each face target in the current video frame;

For the specified face target, when the current video frame meets the preset snapshot interval, the face target image with the best image quality in the preset snapshot interval is uploaded to the comparison system as the captured image of the specified face target.

Optionally, when the current video frame meets the preset capture interval for the specified face target, the face target image with the best image quality in the preset capture interval is uploaded to the comparison system as the captured image of the specified face target Before the steps, the method also includes:

Use the preset target tracking algorithm to associate the target frame between the current video frame and the previous video frame to generate target trajectories corresponding to different facial targets;

According to the target trajectory corresponding to each face target, the same face target in the current video frame and the previous video frame is determined.

Optionally, the method of determining the image quality includes:

Obtain the face target quality parameters of the face target image;

According to the face target quality parameter, the image quality of the face target image is determined.

Optionally, if the preset snapshot interval is 1, uploading the face target image with the best image quality within the preset snapshot interval as the captured image of the specified face target to the comparison system includes:

Upload the face target image of the specified face target in the current video frame to the comparison system as the captured image of the specified face target.

Optionally, the step of uploading the face target image with the best image quality in the preset capture interval as the captured image of the specified face target to the comparison system includes:

Cache the face image with the best image quality and the best image quality for the specified face target within the preset snapshot interval;

Determine whether the image quality of the currently cached face target image is better than the image quality of each previously cached face target image;

If it is better, the currently cached face target image is uploaded to the comparison system as a captured image of the specified face target.

Cache the face target image with the best image quality in the preset snapshot interval as the snapshot image to the cache area;

After determining that the capture of the specified face target is completed, perform video compression on each captured image in the buffer area to obtain the compressed video;

Upload the compressed video to the comparison system.

In a second aspect, an embodiment of the present application provides a surveillance camera, including a surveillance camera, a processor, and a memory, where,

Surveillance camera, used to collect the current video frame;

Memory, used to store computer programs;

The processor, when used to execute the computer program stored on the memory, implements the following steps:

Optionally, when the processor executes the computer program stored on the memory, the following steps are also implemented:

Obtain the face target quality parameters of the face target image;

Optionally, if the preset capture interval is 1, the processor implements the step of uploading the face target image with the best image quality in the preset capture interval as the captured image of the specified face target to the comparison system. Implement the following steps:

Optionally, when the processor implements the step of uploading the face target image with the best image quality within the preset capture interval as the captured image of the specified face target to the comparison system, the specific steps are as follows:

Upload the compressed video to the comparison system.

In a third aspect, an embodiment of the present application provides a machine-readable storage medium in which a computer program is stored. When the computer program is executed by a processor, the image capture method provided in the first aspect of the embodiment of the present application is implemented .

According to a fourth aspect, an embodiment of the present application provides an application program for execution at runtime: the image capturing method provided in the first aspect of the embodiment of the present application.

In a fifth aspect, an embodiment of the present application provides a monitoring system, including a monitoring camera and a comparison system;

Surveillance camera, used to collect the current video frame; perform face target detection on the current video frame to determine the face target image of each face target in the current video frame; for the specified face target, the preset capture interval is met in the current video frame , The face target image with the best image quality within the preset capture interval is uploaded to the comparison system as the captured image of the specified face target;

The comparison system is used to compare and alarm the captured images.

An image capturing method, a monitoring camera and a monitoring system provided by the embodiments of the present application, the monitoring camera collects the current video frame, performs face target detection on the current video frame, and determines the face target image of each face target in the current video frame For the specified face target, when the current video frame meets the preset capture interval, the face target image with the best image quality in the preset capture interval is uploaded to the comparison system as the captured image of the specified face target. When the current video is collected, face target detection is performed on the current video frame, and when the current video frame meets the preset capture interval, for the specified face target, the face image with the best image quality within the preset capture interval Uploading the captured image of the specified face target to the comparison system is equivalent to grouping the video frame sequence for detecting the specified face target, which can be used to capture the specified face target in different capture intervals respectively. The image quality is optimal within the capture interval, to ensure that the surveillance camera uploads multiple capture images with higher image quality to the comparison system, and these capture images are not concentrated in a certain period of time because they are in different time intervals. It has a high degree of richness, which ensures the recognizability of face targets, and thus ensures that the comparison results of the comparison system have a high accuracy.

BRIEF DESCRIPTION

In order to more clearly explain the embodiments of the present application and the technical solutions of the prior art, the following briefly introduces the drawings required in the embodiments and the prior art. Obviously, the drawings in the following description are only For some embodiments of the application, for those of ordinary skill in the art, without paying any creative labor, other drawings may be obtained based on these drawings.

FIG. 1 is a schematic flowchart of an image capture method according to an embodiment of this application;

2 is a schematic diagram of a snapshot effect of an embodiment of the present application;

3 is a schematic diagram of a snapshot effect of another embodiment of the present application;

4 is a schematic diagram of a snapshot effect according to another embodiment of the application;

5 is a schematic diagram of a snapshot effect of yet another embodiment of the present application;

6 is a schematic diagram of a process of capturing and compressing a face target image according to an embodiment of the present application;

7 is a schematic structural diagram of a surveillance camera according to an embodiment of this application;

8 is a schematic structural diagram of a monitoring system according to an embodiment of the present application.

detailed description

In order to make the purpose, technical solutions and advantages of the present application more clear, the following describes the present application in further detail with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

In practical applications, the monitoring system mainly captures, compares and alarms face targets. The monitoring system includes a surveillance camera and a comparison system. The comparison system can be a background server, which is mainly used to realize feature extraction, face comparison and alarm functions.

In order to ensure that the comparison results of the comparison system have high accuracy, the embodiments of the present application provide an image capture method, a monitoring camera, a machine-readable storage medium, and a monitoring system.

In the following, the image capture method provided by the embodiments of the present application is first introduced.

The execution subject of the image capture method provided in the embodiments of the present application may be a surveillance camera (for example, a smart camera, a network camera, etc.) in the surveillance system, and the surveillance camera may include at least a surveillance camera and a processor equipped with a core processing chip . The method for implementing the image capturing method provided by the embodiments of the present application may be at least one method of software, hardware circuits, and logic circuits provided in the monitoring camera.

As shown in FIG. 1, an image capture method provided by an embodiment of the present application may include the following steps:

S101: Collect the current video frame.

Surveillance cameras can be installed in all corners of the city, for example, community entrances, intersections, parks, stadiums, etc. Here, there is no specific requirement for the specific location, angle and resolution of the surveillance camera, which can meet the coverage. As long as possible, the requirements for capturing the face target clearly. The surveillance camera can shoot the surveillance scene in real time to obtain video data of the surveillance scene.

The surveillance camera can shoot the surveillance scene in real time to obtain the video data of the surveillance scene, and the video data includes the video frames of each frame and the time stamp of each video frame collected. In order to ensure the real-time nature of the face target image, the current video frame collected needs to be processed.

S102: Perform face target detection on the current video frame to determine the face target image of each face target in the current video frame.

You can use the preset target detection algorithm to detect the face of the current video frame. The preset target detection algorithm can be a traditional feature matching algorithm, which can determine the current video frame by face features such as eyes, nose, mouth, ears, etc. Whether the target in is a face target, if it is a face target, a certain area around the face target is divided into a face target frame, and the image in the face target frame or the image within a certain range of the face target frame is a person Face target image; the preset target detection algorithm can also be a more popular intelligent detection algorithm, such as deep neural network. The network model of deep neural network can be obtained by training a large number of face images, by inputting the current video frame into the depth The neural network can obtain the interest area of the face target in the current video frame, and the image in the interest area of the face target or the image within a certain range of the interest area of the face target is the face target image. Of course, other methods that can detect the face target in the video frame also belong to the protection scope of the embodiments of the present application, and details are not repeated here.

S103: For the specified face target, when the current video frame meets the preset capture interval, upload the face target image with the best image quality in the preset capture interval as the captured image of the specified face target to the comparison system.

The specified face target can be any face target previously detected. For a specified face target, the first video frame in which the specified face target is detected can be regarded as the start frame, and every subsequent frame is collected , The frame number is superimposed once. If the number of frames in the current video frame satisfies the preset snapshot interval, it means that the condition for grouping to determine the captured image is reached. The face target with the best image quality within this preset snapshot interval needs to be The image is taken as the captured image of the specified face target. For example, for the face target A, the face target A is detected for the first time in the fifth frame, then the fifth frame is recorded as the start frame of the face target A, and the current video frame is the 20th frame, that is, the person is detected In the fifteenth frame after face target A, if the preset capture interval is 15 frames, the face target image with the best image quality in these 15 frames can be used as the capture image of face target A.

When the current video frame meets the preset snapshot interval, it is equivalent to grouping the video frame sequence that detects the specified face target. The video frames in a snapshot interval are divided into a group, and the preset snapshot interval is the number of video frames. For example, if the snapshot interval is set to 10 frames, as shown in FIG. 2, the shaded part indicates that the image quality of the face target image in the video frame is the best. Then the captured image includes: the corresponding face target specified in frames 1-10 The face target image with the highest image quality (the face target image corresponding to the face target in frame 6), and the face image with the highest image quality corresponding to the face target in frames 11-20 (frame 19 The face target image corresponding to the face target in frame), the face target image with the highest image quality corresponding to the face target in frames 21-30 (the face target image corresponding to the face target in frame 23), The face image with the highest image quality corresponding to the face target in frames 31-40 (the face target image corresponding to the face target in frame 40), and the face image corresponding to the face target in frames 41-50 The face image with the highest image quality (Face target image corresponding to the face target in frame 47). The preset snapshot interval may be fixed or variable. For example, in the first three periods, the preset snapshot interval may be set to 15 frames, and in the latter period, the preset snapshot interval may be set to 25 frames, and so on.

For the current video frame collected, the image quality analysis of the face target images of each face target in the video is required to determine the face target image with the best image quality. Optionally, determine the image quality The specific way can be:

Obtain the face target quality parameters of the face target image; determine the image quality of the face target image based on the face target quality parameters.

There are many factors that affect the image quality of the face target image, such as the degree to which the face target is blocked in the face target image, the imaging clarity of the face target in the face target image, and the pose of the face target in the face target image and many more. The less the face target is occluded in the face target image, the higher the image quality, the clearer the face target is imaged in the face target image, the higher the image quality, and the more positive the face target in the face target image More often the image quality is higher and so on. Of course, there are many factors that affect the image quality of the face target image, such as contrast, brightness, etc., which are not listed here.

You can consider the above factors comprehensively and assign a certain image quality score value to the face target image by weighting; you can also consider only one or some of the factors set by the analysis algorithm to assign a certain image quality score to the face target image value. The preset quality analysis algorithm can also assign, for example, good, good, medium, and poor evaluation results to the face target image.

The face target quality parameter refers to the parameter that affects the image quality of the face target image, mainly including the face target's posture, degree of occlusion, and imaging clarity. When performing image quality analysis on the face target image, you can obtain the person The facial target quality parameters such as the posture information, degree of occlusion and sharpness of the facial target are based on the comprehensive consideration of the effects of different facial target quality parameters on the image quality to obtain the image quality of the facial target image. For example, in the current video frame, in the face target image in the face target frame, the face target A is completely frontal, the face is occluded by 1/10, and the definition is very high, the image quality of the face target image can be determined For superiority, or, to quantify the image quality, assign an image quality score of 9.

The way to determine the face target image with the best image quality may be to cache the determined face target image and image quality, and after collecting the current video frame, compare the face target image in the current video frame with the cached The image quality of the face target image. If the image quality of the face target image in the current video frame is better, the face target image in the current video frame is cached, overwriting the original face target image cached. When the video frame meets the preset capture interval, the face target image with the best image quality can be directly determined as the captured image; of course, the way to determine the face target image with the best image quality can also be every time a face is detected The target image caches the face target image and image quality until it stops when the current video frame meets the preset capture interval, and then compares the whole image to find the face target image with the best image quality as the captured image. There is no specific limit here.

You can immediately upload the captured image to the comparison system to ensure the real-time performance of the comparison system; you can also cache the captured image and select part or all of the captured image from the cache area to upload to the comparison after the face target disappears System to ensure the accuracy of the comparison system.

In order to reduce the transmission pressure between the surveillance camera and the comparison system, the snapshot camera can select a preset number of snapshot images from the snapshot images determined in each preset snapshot interval to upload, instead of uploading all the determined snapshot images All upload, the preset number can be one or more. In practical applications, if only one image is captured, and the image quality of the captured image is not particularly high, it may not be able to support the high accuracy of the comparison result of the comparison system, or even if the image quality of the captured image is high Will also affect the comparison result because the richness of the face target is too poor. Therefore, the preset number is usually set to multiple. For the selection of multiple captured images, it can be multiple captured images with the highest image quality. It may be a plurality of captured images determined first, and no specific limitation is made here.

Optionally, before S103, the image capturing method provided by the embodiment of the present application may further perform the following steps:

According to the target trajectory corresponding to each face target, the same face target in the current video and the previous video frame is determined.

For the tracking of the same face target in different video frames, it can be achieved by using a preset target tracking algorithm. The preset target tracking algorithm can be a currently popular intelligent algorithm by targeting the face target in each video frame Frame association generates target trajectories corresponding to different face targets. The same target trajectory corresponds to the same face target. In this way, the same face target in the current video frame and the previous video frame can be determined according to the target trajectory.

Of course, for tracking the same face target in different video frames, it can also be realized by matching the target frame after performing face target detection on the current video frame and the previous video frame. If the face target frame can be matched, it means that the two face target frames correspond to the same face target.

Optionally, if the preset snapshot interval is 1, S103 may specifically be:

If the preset snapshot interval is 1 frame, it means that the capture of the surveillance camera is actually a full-frame snapshot, that is, for the specified face target, as long as the face target image of the specified face target is detected in the current video frame, it will be The target image of the face is taken as a captured image, and the effect picture of the captured image is shown in FIG. 3.

Optionally, S103 may specifically be:

If the preset snapshot interval is greater than 1 frame, each preset snapshot interval needs to be captured. Here, the snapshot can be captured in increasing quality to ensure that the captured image uploaded to the comparison system is the best image quality of all captured images. Several good ones. For the specified face target, each time the face target image with the best image quality is determined, it is cached and the optimal image quality is cached. If the image quality of the currently cached face target image is better than the previously cached faces Only the image quality of the target image is taken as the captured image of the face image currently cached, otherwise it will not be captured.

As shown in FIG. 4, the face image with the highest image quality corresponding to the specified face target in frames 11-10 is the face target image corresponding to the specified face target in frame 6, then the person in frame 6 The face target image is taken as a captured image; the face image with the highest image quality corresponding to the specified face target in frames 11-20 is the face target image corresponding to the specified face target in frame 19, but specified in frame 19 The image quality of the face target image corresponding to the face target is inferior to the face target image corresponding to the specified face target in frame 6, so no snapshot is taken; the person with the highest image quality corresponding to the specified face target in frames 21-30 The face target image is the face target image corresponding to the face target specified in frame 23, and the image quality of the face target image corresponding to the face target specified in frame 23 is higher than the face target specified in frame 6 and frame 19 The corresponding face target image is better, so the face target image in frame 23 is taken as the captured image; the face image with the highest image quality corresponding to the specified face target in frames 31-40 is in frame 40 The corresponding face target Face target image, and the image quality of the face target image corresponding to the face target specified in frame 40 is better than the face target image corresponding to the face target specified in frames 6, 19, and 23; therefore, Use the face target image in frame 40 as the captured image; the face image with the highest image quality corresponding to the specified face target in frames 41-50 is the face target image corresponding to the specified face target in frame 48, And the image quality of the face target image corresponding to the specified face target in frame 48 is inferior to the face target image corresponding to the specified face target in frames 6, 19, 23, and 40, so it is not captured .

According to actual needs, there are also surveillance cameras that only need to upload a face target image with the best image quality as a captured image to the comparison system, so that the image quality of all face target images can be compared and the best image quality can be selected from It is better to upload as a captured image, or to use an overlay cache method. Always cache a face target image with better image quality. When the face target disappears, the cached face image is the face image with the best image quality. , Upload the face image as a captured image. The effect picture of the snapshot is shown in FIG. 5. The image quality of the face target image corresponding to the specified face target in the 40th frame is the best. You only need to determine the face target image in the 40th frame as the captured image. Compare systems.

Optionally, S103 may specifically be:

Upload the compressed video to the comparison system.

After the captured image is determined, the captured image can be cached in the cache area, and then the entire compression upload is performed. The process is shown in FIG. 6. When the capture of the specified face target is completed, that is, when the specified face target is determined to disappear Afterwards, all cached captured images are compressed. Although the same face target has different postures at different moments, for a video, most areas such as the background are the same, and the same face target has correlation between different video frames. Therefore, by compressing multiple captured images with video, compared with JPEG (Joint Photographic Experts Group) encoding and compression for each image in the traditional capture method, it can save more bandwidth and reduce transmission and Storage cost.

The way to determine the end of the capture of the face target may be that the matching degree of the continuous multiple frames to the face target is very low, then the end of the capture of the face target may be determined, or the target tracking algorithm may be used to determine the loss of the face target , That is, a face target cannot be tracked in multiple consecutive frames, and it can be determined that the capture of the face target ends.

Applying this embodiment, the monitoring camera collects the current video frame and performs face target detection on the current video frame to determine the face target image of each face target in the current video frame. For the specified face target, the current video frame meets the pre- When the snapshot interval is set, the face target image with the best image quality in the preset snapshot interval is uploaded to the comparison system as the captured image of the specified face target. When the current video is collected, face target detection is performed on the current video frame, and when the current video frame meets the preset capture interval, for the specified face target, the face image with the best image quality within the preset capture interval Uploading the captured image of the specified face target to the comparison system is equivalent to grouping the video frame sequence for detecting the specified face target, which can be used to capture the specified face target in different capture intervals respectively. The image quality is optimal within the capture interval, to ensure that the surveillance camera uploads multiple capture images with higher image quality to the comparison system, and these capture images are not concentrated in a certain period of time because they are in different time intervals. It has a high degree of richness, which ensures the recognizability of face targets, and thus ensures that the comparison results of the comparison system have a high accuracy.

Corresponding to the above method embodiment, an embodiment of the present application provides a surveillance camera. As shown in FIG. 7, it includes a surveillance camera 701, a processor 702, and a memory 703, where,

Surveillance camera 701, used to collect the current video frame;

Memory 703, used to store computer programs;

When the processor 702 is used to execute the computer program stored on the memory 703, the following steps are implemented:

Optionally, when the processor 702 executes the computer program stored on the memory, the following steps may also be implemented:

Obtain the face target quality parameters of the face target image;

Optionally, if the preset snapshot interval is 1, the processor 702, when implementing the step of uploading the face target image with the best image quality in the preset snapshot interval as the captured image of the specified face target to the comparison system, The specific steps can be achieved as follows:

Optionally, when the processor 702 implements the step of uploading the face target image with the best image quality within the preset capture interval as the captured image of the specified face target to the comparison system, the following steps may be specifically implemented:

Upload the compressed video to the comparison system.

The above memory may include RAM (Random Access Memory, random access memory), or may include NVM (Non-Volatile Memory, non-volatile memory), for example, at least one disk memory. Optionally, the memory may also be at least one storage device located away from the processor.

The above processor may be a general-purpose processor, including CPU (Central Processing Unit), NP (Network Processor), etc .; it may also be DSP (Digital Signal Processing, digital signal processor), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

Data can be transmitted between the surveillance camera 701, the processor 702, and the memory 703 through a wired connection or a wireless connection, and the surveillance camera can communicate with the comparison system through a wired communication interface or a wireless communication interface. 7 is only an example of data transmission between the surveillance camera 701, the processor 702, and the memory 703 through the bus, and is not intended as a limitation of a specific connection method.

In this embodiment, the processor of the surveillance camera can read the computer program stored in the memory and run the computer program to realize that the surveillance camera collects the current video frame and performs face target detection on the current video frame to determine The face target image of each face target in the current video frame, for the specified face target, when the current video frame meets the preset snapshot interval, the face target image with the best image quality in the preset snapshot interval is used as the designated person The captured image of the face target is uploaded to the comparison system. When the current video is collected, face target detection is performed on the current video frame, and when the current video frame meets the preset capture interval, for the specified face target, the face image with the best image quality within the preset capture interval Uploading the captured image of the specified face target to the comparison system is equivalent to grouping the video frame sequence for detecting the specified face target, which can be used to capture the specified face target in different capture intervals respectively. The image quality is optimal within the capture interval, to ensure that the surveillance camera uploads multiple capture images with higher image quality to the comparison system, and these capture images are not concentrated in a certain period of time because they are in different time intervals. It has a high degree of richness, which ensures the recognizability of face targets, and thus ensures that the comparison results of the comparison system have a high accuracy.

In addition, embodiments of the present application also provide a machine-readable storage medium, and the machine-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the image capturing method provided by the embodiment of the present application All steps.

In this embodiment, the machine-readable storage medium stores a computer program that executes the image capture method provided by the embodiment of the present application at runtime, so it can be achieved that the surveillance camera performs face targeting on the current video frame by collecting the current video frame Detection to determine the face target image of each face target in the current video frame. For the specified face target, when the current video frame meets the preset snapshot interval, the face target image with the best image quality within the preset snapshot interval is used as The captured image of the designated face target is uploaded to the comparison system. When the current video is collected, face target detection is performed on the current video frame, and when the current video frame meets the preset capture interval, for the specified face target, the face image with the best image quality within the preset capture interval Uploading the captured image of the specified face target to the comparison system is equivalent to grouping the video frame sequence for detecting the specified face target, which can be used to capture the specified face target in different capture intervals respectively. The image quality is optimal within the capture interval, to ensure that the surveillance camera uploads multiple capture images with higher image quality to the comparison system, and these capture images are not concentrated in a certain period of time because they are in different time intervals. It has a high degree of richness, which ensures the recognizability of face targets, and thus ensures that the comparison results of the comparison system have a high accuracy.

An embodiment of the present application also provides an application program for executing at run time: all steps of the image capturing method provided by the embodiment of the present application.

In the embodiment of the present application, the application program executes the image capturing method provided in the embodiment of the present application when it is running, so that it can realize that the surveillance camera performs face target detection on the current video frame by collecting the current video frame to determine the current video frame. The face target image of each face target, for the specified face target, when the current video frame meets the preset capture interval, the face target image with the best image quality within the preset capture interval is used as the snapshot of the specified face target The image is uploaded to the comparison system. When the current video is collected, face target detection is performed on the current video frame, and when the current video frame meets the preset capture interval, for the specified face target, the face image with the best image quality within the preset capture interval Uploading the captured image of the specified face target to the comparison system is equivalent to grouping the video frame sequence for detecting the specified face target, which can be used to capture the specified face target in different capture intervals respectively. The image quality is optimal within the capture interval, to ensure that the surveillance camera uploads multiple capture images with higher image quality to the comparison system, and these capture images are not concentrated in a certain period of time because they are in different time intervals. It has a high degree of richness, which ensures the recognizability of face targets, and thus ensures that the comparison results of the comparison system have a high accuracy.

An embodiment of the present application also provides a monitoring system. As shown in FIG. 8, the monitoring system may include a monitoring camera 810 and a comparison system 820;

Surveillance camera 810, used to collect the current video frame; perform face target detection on the current video frame to determine the face target image of each face target in the current video frame; for the specified face target, the current video frame meets the preset snapshot At intervals, the face target image with the best image quality within the preset capture interval is uploaded to the comparison system as the captured image of the specified face target;

The comparison system 820 is used for comparing and alarming the captured images.

The monitoring camera 810 can also be used to implement all the steps provided in the above method embodiments, which will not be repeated here.

The comparison system 820 compares and alarms the captured image, which may include: extracting the captured image feature, and comparing the extracted facial features with the facial features in the blacklist, if the similarity of the comparison is greater than a certain Threshold, then alarm.

For the embodiments of the surveillance camera, the machine-readable storage medium, the application program, and the surveillance system, since the content of the methods involved is basically similar to the foregoing method embodiments, the description is relatively simple. Partial instructions are sufficient.

It should be noted that in this article, relational 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 that these entities or operations There is any such actual relationship or order. Moreover, the terms "include", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also those not explicitly listed Or other elements that are inherent to this process, method, article, or equipment. Without more restrictions, the element defined by the sentence "include one ..." does not exclude that there are other identical elements in the process, method, article or equipment that includes the element.

The embodiments in this specification are described in a related manner. The same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, monitoring camera, machine-readable storage medium, application program, and monitoring system embodiments, since they are basically similar to the method embodiments, the description is relatively simple. For the related parts, refer to the section of the method embodiments. can.

The above are only the preferred embodiments of this application and are not intended to limit this application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application should be included in this application Within the scope of protection.

Claims

An image capture method, characterized in that the method includes:

Collect the current video frame;

Perform face target detection on the current video frame to determine the face target image of each face target in the current video frame;

For the specified face target, when the current video frame meets the preset snapshot interval, upload the face target image with the best image quality in the preset snapshot interval as the captured image of the specified face target to the comparison system.
The method according to claim 1, wherein, for the specified face target, when the current video frame satisfies a preset snapshot interval, the face with the best image quality in the preset snapshot interval is selected Before the target image is uploaded to the comparison system as a captured image of the specified face target, the method further includes:

Use a preset target tracking algorithm to associate the target frame with the current video frame and the previous video frame to generate target trajectories corresponding to different face targets;

The same face target in the current video frame and the previous video frame is determined according to the target trajectory corresponding to each face target.
The method according to claim 1 or 2, wherein the method for determining the image quality includes:

Acquiring the face target quality parameters of the face target image;

The image quality of the face target image is determined according to the face target quality parameter.
The method according to claim 1, wherein if the preset snapshot interval is 1, the face target image with the best image quality in the preset snapshot interval is used as the specified face target The captured images are uploaded to the comparison system, including:

Upload the face target image of the specified face target in the current video frame to the comparison system as a captured image of the specified face target.
The method according to claim 1, wherein the uploading of the face target image with the best image quality in the preset capture interval as the captured image of the specified face target to the comparison system includes:

Cache the face image with the best image quality and the best image quality for the specified face target within the preset snapshot interval;

Determine whether the image quality of the currently cached face target image is better than the image quality of each previously cached face target image;

If it is better, the current cached face target image is uploaded to the comparison system as a captured image of the specified face target.
The method according to claim 1, wherein the uploading of the face target image with the best image quality in the preset capture interval as the captured image of the specified face target to the comparison system includes:

Cache the face target image with the best image quality in the preset capture interval as the captured image in the cache area;

After determining that the capture of the designated face target is completed, perform video compression on each captured image in the buffer area to obtain a compressed video;

Upload the compressed video to the comparison system.
A surveillance camera is characterized by comprising a surveillance camera, a processor and a memory, wherein,

The monitoring camera is used to collect the current video frame;

The memory is used to store computer programs;

When the processor is used to execute the computer program stored on the memory, the following steps are implemented:

Perform face target detection on the current video frame to determine the face target image of each face target in the current video frame;

For the specified face target, when the current video frame meets the preset snapshot interval, upload the face target image with the best image quality in the preset snapshot interval as the captured image of the specified face target to the comparison system.
The surveillance camera according to claim 7, wherein when the processor executes the computer program stored on the memory, the following steps are further implemented:

Use a preset target tracking algorithm to associate the target frame with the current video frame and the previous video frame to generate target trajectories corresponding to different face targets;

The same face target in the current video frame and the previous video frame is determined according to the target trajectory corresponding to each face target.
The surveillance camera according to claim 7 or 8, wherein when the processor executes the computer program stored on the memory, the following steps are further implemented:

Acquiring the face target quality parameters of the face target image;

The image quality of the face target image is determined according to the face target quality parameter.
The surveillance camera according to claim 7, wherein, if the preset snapshot interval is 1, the processor is implementing the face target image with the best image quality in the preset snapshot interval As the step of uploading the captured image of the designated face target to the comparison system, the following steps are specifically implemented:

Upload the face target image of the specified face target in the current video frame to the comparison system as a captured image of the specified face target.
The surveillance camera according to claim 7, wherein the processor is configured to upload the face target image with the best image quality in the preset capture interval as the captured image of the specified face target When comparing the steps of the system, the specific steps are as follows:

Cache the face image with the best image quality and the best image quality for the specified face target within the preset snapshot interval;

Determine whether the image quality of the currently cached face target image is better than the image quality of each previously cached face target image;

If it is better, the current cached face target image is uploaded to the comparison system as a captured image of the specified face target.
The surveillance camera according to claim 7, wherein the processor is configured to upload the face target image with the best image quality in the preset capture interval as the captured image of the specified face target When comparing the steps of the system, the specific steps are as follows:

Cache the face target image with the best image quality in the preset capture interval as the captured image in the cache area;

After determining that the capture of the designated face target is completed, perform video compression on each captured image in the buffer area to obtain a compressed video;

Upload the compressed video to the comparison system.
An application program, which is used to execute at runtime: the method according to any one of claims 1-6.
A monitoring system, characterized in that it includes a monitoring camera and a comparison system;

The surveillance camera is used to collect the current video frame; perform face target detection on the current video frame to determine the face target image of each face target in the current video frame; for the specified face target, in the When the current video frame satisfies the preset snapshot interval, upload the face target image with the best image quality in the preset snapshot interval as the captured image of the specified face target to the comparison system;

The comparison system is used for comparing and alarming the captured images.