Road image extraction method and device in intelligent video monitoring

Abstract

The invention relates to the intelligent video monitoring field, in particular to a road image extraction technique. The invention provides a road image extraction method and a device used in an intelligent video monitoring system and used for exactly defining the movement zone of a moving object in a monitored site. The road image extraction method includes the steps that: input images are obtained from a video flow and a background image is generated according to the obtained input images; movement pixel points of input images obtained from the video flow are inspected according to the background image to generate moving object images; the frames of the generated moving object images are counted, and times of each pixel point belonging to the movement pixel points are respectively counted according to the movement pixel points of each frame of the moving object image; when the generated moving object image frames reach a set frame quantity, the pixel points with counted times greater than a first threshold value are defined to form the movement zone of the moving object, and the road images are generated according to background image and the defined movement zone; wherein, the background image can be updated according to the moving object images and the input images.

Description

technical field [0001] The invention relates to the field of intelligent video monitoring, in particular to a road surface image extraction technology. Background technique [0002] Intelligent Video Surveillance (IVS) is a monitoring method based on computer vision technology to analyze video images of surveillance scenes, extract key information in surveillance scenes, and generate corresponding events and alarms. Intelligent video monitoring technology uses the powerful data processing function of the computer to analyze the massive data in the video image at high speed, filter out the information that the monitor does not care about, and provide useful key information for the monitor. [0003] Intelligent video surveillance is a high-end video surveillance application based on digital and networked video surveillance. The background image in the monitoring scene refers to the image composed of the static background. For example, a frame image extracted from the video st...

AI Technical Summary

Problems solved by technology

[0008] In the prior art, the background image is mainly extracted through a simple differential operation on a series of input images, the extraction efficiency is low, the accuracy of the extracted background image is poor, and it is impossible to accurately distinguish the moving area and the non-moving area of ​​the moving target, and even Errors may occur in complex background environments

For example, for intelligent traffic monitoring, the existing technology mainly uses various commonly used motion detection methods to estimate the background image and then detect the vehicles on the road, and then perform subsequent operations such as lane line detection, vehicle speed estimation, and vehicle size estimation. , these methods are more effective when there is only one lane in the monitoring field of view, but if there are multiple lanes in the field of view, the efficiency of these methods will be relatively low, and even errors will occur

Images