阅读说明：本技术 一种智能监控视频处理方法 (Intelligent monitoring video processing method ) 是由 马培娜 张�林 于登昌 韩克强 杨守斌 王成锐 于 2019-08-20 设计创作，主要内容包括：本发明公开了一种智能监控视频处理方法,包括如下步骤：(1)视频预处理：将视频转化为静态图像,然后过滤视频图像中的噪点,同时对整体图像的数据做偏移调整,使像素灰度分布均匀；(2)前景提取：利用静止的空场景图像进行背景重构,然后根据预设区域,增加系统空间标识,再对新的图像进行背景减除,提取差异较大的像素区域,做为活动前景；(3)行为跟踪：对提取出来的活动前景进行卷积处理,取前景轮廓；然后提取特征点,记录活动轨迹；(4)结果分析：分析最终结果,并通过各业务要求对各类数据相应存储,以满足各业务功能要求。本发明所公开的方法能够更加快速有效的协助安全人员处理危机,最大限度的降低误报和漏报现象。(The invention discloses an intelligent monitoring video processing method, which comprises the following steps: (1) video preprocessing: converting a video into a static image, filtering noise in the video image, and performing offset adjustment on data of the whole image to ensure that the gray level of pixels is uniformly distributed; (2) and (3) foreground extraction: utilizing a static empty scene image to reconstruct a background, then adding a system space identifier according to a preset area, then carrying out background subtraction on a new image, and extracting a pixel area with larger difference as an active foreground; (3) behavior tracking: performing convolution processing on the extracted active foreground, and taking a foreground outline; then extracting characteristic points and recording an activity track; (4) and (4) analyzing results: and analyzing the final result, and correspondingly storing various data according to various service requirements so as to meet the functional requirements of various services. The method disclosed by the invention can be used for more quickly and effectively assisting safety personnel in handling crisis, and the phenomena of false alarm and missing report are reduced to the maximum extent.)

1. An intelligent monitoring video processing method is characterized by comprising the following steps:

(1) video preprocessing: converting a video into a static image, filtering noise in the video image, and performing offset adjustment on data of the whole image to ensure that the gray level of pixels is uniformly distributed;

(2) and (3) foreground extraction: utilizing a static empty scene image to reconstruct a background, then adding a system space identifier according to a preset area, then carrying out background subtraction on a new image, and extracting a pixel area with larger difference as an active foreground;

(3) behavior tracking: performing convolution processing on the extracted active foreground to realize edge detection and obtain a foreground outline; then extracting characteristic points, simulating and representing the motion and morphological change of the foreground contour by using a geometric model, and recording a motion track;

(4) and (4) analyzing results: and comparing and analyzing final results through cross comparison, background analysis, behavior classification and comprehensive data of motion trail and region intensity, and correspondingly storing various data through various business requirements so as to meet the functional requirements of various businesses.

2. The intelligent surveillance video processing method according to claim 1, wherein the specific method for converting the video into the still image in step (1) is: the method comprises the steps of firstly obtaining the number of frames per second of a video by adopting an equal-interval video capture mode, and then randomly extracting one frame of the frames per second for video capture, thereby obtaining a static image.

3. The intelligent surveillance video processing method according to claim 1, wherein the step (1) of performing offset adjustment on the data of the whole image comprises: the gray stretching technology is adopted to convert the separated gray into a more concentrated degree, the gray enhancement algorithm is adopted in the processing kernel to enhance the image contrast by stretching the pixel intensity distribution range, and then the median filtering method is adopted to carry out image smoothing treatment in a mode of expanding the image function matrix and filling the matrix edge.

4. The intelligent surveillance video processing method according to claim 2, wherein the specific method of step (2) is: a series of frames which are captured and optimized in video preprocessing are subjected to average background technology, and the average pixel value in the frames is taken to represent the background; then adding a certain threshold range to the average pixel values to form a background model; in the newly added image, if the pixel of the corresponding position exceeds the threshold range of the pixel of the corresponding position in the background model, the pixel is taken as a processing basis, then the space scene model provided by the application function is compared with the image after average background processing to reconstruct the space background and increase the space identification, and then the stability of the pixel point is judged through simple threshold operation by a background subtraction algorithm CNT method; and if the pixel points are stable in continuous frames, the pixel points are considered to be stable, otherwise, the pixel points are unstable, and the stable points are background points in the program.

5. The intelligent surveillance video processing method according to claim 1, wherein the specific method of step (3) is: calculating an approximate value of the gray scale of the image brightness function by using a Sobel algorithm and two groups of 3 x 3 matrix discrete difference operators of Sobel convolution factors, and detecting points with obvious change in the digital image; meanwhile, the system opens a convolution factor updating function, supports the continuous increase of convolution factors, performs convolution operation, and enlarges the difference between the target and the target so as to analyze various behaviors in the image.

6. The intelligent surveillance video processing method according to claim 1, wherein the step (4) employs a descriptive statistical method, specifically including a subtraction method, an averaging method, a minimum neighbor method, a ratio regression method, and a decision tree method.

7. The intelligent surveillance video processing method as claimed in claim 1, wherein the step (4) further comprises a hypothesis testing method, and the parameter testing is testing some main parameters under the condition of known population distribution, and the main parameters comprise mean, percentage, variance and correlation coefficient.

8. The intelligent surveillance video processing method according to claim 1, wherein the step (4) further includes a deep analysis extension method, and when the user has a deep analysis requirement, the judgment is performed based on the assumption conditions defined by the system or the assumption assignment of the analysis parameters required by other analysis methods.

Technical Field

The invention relates to a video processing method, in particular to an intelligent monitoring video processing method.

Background

The intelligent video monitoring system is not long in development time, but greatly contributes to the environment with social security and stability, so various factors need to be considered in specific design. The video monitoring system has the advantages of real-time, recordable and visible performance, and meanwhile, the recorded information amount is large, the application range is wide, and therefore the video monitoring system can play an important role in various places such as security, traffic, production and life and the like.

The intelligent image monitoring system is a product combining various latest technologies of multimedia, image processing, computer and the like, and adopts computer vision processing, pattern recognition, computer graphic image processing technology and the like. It converts the analog video signal into digital signal by using the latest image digital processing technology, and synchronously stores the video signals in the computer hard disk in a data stream mode while displaying multiple paths (1-20 paths) of moving images on the computer display in real time. The monitoring, recording and playback of the video signals are realized on the computer.

The conventional monitoring system often needs manual intervention, real-time monitoring needs personnel to stare at a monitor, search for playback needing the personnel at a little by a little, and the like. When the video is searched, the video is very inconvenient, a large amount of time is consumed, and the labor cost is increased.

Disclosure of Invention

In order to solve the technical problems, the invention provides an intelligent monitoring video processing method, so as to achieve the purposes of realizing one-time input of conditions, automatically alarming and reducing the burden of monitoring personnel.

In order to achieve the purpose, the technical scheme of the invention is as follows:

an intelligent monitoring video processing method comprises the following steps:

(1) video preprocessing: converting a video into a static image, filtering noise in the video image, and performing offset adjustment on data of the whole image to ensure that the gray level of pixels is uniformly distributed;

(2) and (3) foreground extraction: utilizing a static empty scene image to reconstruct a background, then adding a system space identifier according to a preset area, then carrying out background subtraction on a new image, and extracting a pixel area with larger difference as an active foreground;

(3) behavior tracking: performing convolution processing on the extracted active foreground to realize edge detection and obtain a foreground outline; then extracting characteristic points, simulating and representing the motion and morphological change of the foreground contour by using a geometric model, and recording a motion track;

(4) and (4) analyzing results: and comparing and analyzing final results through cross comparison, background analysis, behavior classification and comprehensive data of motion trail and region intensity, and correspondingly storing various data through various business requirements so as to meet the functional requirements of various businesses.

In the above scheme, the specific method for converting the video into the still image in the step (1) is as follows: the method comprises the steps of firstly obtaining the number of frames per second of a video by adopting an equal-interval video capture mode, and then randomly extracting one frame of the frames per second for video capture, thereby obtaining a static image.

In the above scheme, the specific method for performing offset adjustment on the data of the whole image in the step (1) is as follows: the gray stretching technology is adopted to convert the separated gray into a more concentrated degree, the gray enhancement algorithm is adopted in the processing kernel to enhance the image contrast by stretching the pixel intensity distribution range, and then the median filtering method is adopted to carry out image smoothing treatment in a mode of expanding the image function matrix and filling the matrix edge.

In a further technical scheme, the specific method of the step (2) comprises the following steps: a series of frames which are captured and optimized in video preprocessing are subjected to average background technology, and the average pixel value in the frames is taken to represent the background; then adding a certain threshold range to the average pixel values to form a background model; in the newly added image, if the pixel of the corresponding position exceeds the threshold range of the pixel of the corresponding position in the background model, the pixel is taken as a processing basis, then the space scene model provided by the application function is compared with the image after average background processing to reconstruct the space background and increase the space identification, and then the stability of the pixel point is judged through simple threshold operation by a background subtraction algorithm CNT method; and if the pixel points are stable in continuous frames, the pixel points are considered to be stable, otherwise, the pixel points are unstable, and the stable points are background points in the program.

In the above scheme, the specific method of step (3) is as follows: calculating an approximate value of the gray scale of the image brightness function by using a Sobel algorithm and two groups of 3 x 3 matrix discrete difference operators of Sobel convolution factors, and detecting points with obvious change in the digital image; meanwhile, the system opens a convolution factor updating function, supports the continuous increase of convolution factors, performs convolution operation, and enlarges the difference between the target and the target so as to analyze various behaviors in the image.

In the above scheme, the step (4) adopts a descriptive statistical method, which specifically includes a removal method, an averaging method, a minimum neighbor method, a ratio regression method, and a decision tree method.

In a further technical scheme, the step (4) further comprises a hypothesis testing method, and the parameter testing is performed on some main parameters under the condition of known overall distribution, wherein the main parameters comprise a mean value, a percentage, a variance and a correlation coefficient.

In a further technical solution, the step (4) further includes a deep analysis extension method, and when the user has a deep analysis requirement, the judgment is performed on the basis of a hypothesis condition defined by the system or a hypothesis assignment performed on analysis parameters required by other analysis methods.

Through the technical scheme, the intelligent monitoring video processing method provided by the invention is based on digital and networked video monitoring, a user can set certain specific rules, the system identifies different objects and simultaneously identifies whether target behaviors accord with the rules, once the abnormal condition in a monitoring picture is found, the system can send an alarm and provide useful information in a fastest and optimal mode, the time for finding specific information from the past monitoring information is greatly reduced, so that safety personnel can be effectively assisted to process crises, and the phenomena of false alarm and false alarm omission are reduced to the maximum extent.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below.

The invention provides an intelligent monitoring video processing method, which needs relatively simple equipment and specifically comprises the following steps:

high definition camera (or hard disk video recorder): and collecting real-time video data, and sending the real-time video data to a server in real time or periodically to support products of various mainstream models in the market.

A monitor: the display device displays data acquired by a high-definition camera (or a hard disk video recorder) in real time, supports all VGA, HDMI and other interfaces and supports RTMP and GB/T28181 standard protocols.

A server: the method receives video data collected by a high-definition camera (or a hard disk video recorder) and displays the video data to a monitor, and under the conditions of few terminals and low requirement on operation speed, the method can be replaced by a common PC.

Client (optional): real-time video data can be displayed according to authority setting, previous video data can be searched, angles of all cameras can be adjusted, and the method can be used by a common PC.