Improved consecutive frame difference method

Abstract

The invention discloses an improved consecutive frame difference method. The method includes steps of S1, setting a slide window, reading two adjacent frames of images of a video stream from the firstframe and the second frame and ending the algorithm if the last frame is reached; S2, performing filtering treatment on the two frames of images through a high frequency filter; S3, calculating the mean value and the variance by utilizing an integer image method and an integrated square method; S4, calculating the difference between the mean value and the square difference; S5, comparing the difference with a threshold value and performing judgment; S6, judging whether the last frame is reach or not, if not, repeating the above process, if yes, ending the process. The invention discloses theimproved consecutive frame different method. According to the invention, by utilizing the integer image method and the integrated square method, calculation rate and precision increase are realized. The method needs not overall threshold value, high in adaptability and small in calculation amount. Shortcomings of image blurring and insufficient identifiability in a poor environment when a traditional frame difference method is adopted are made up. The method has characteristics of being fully digital and intelligent and being reliable in monitoring a moving target in real time.

Description

technical field [0001] In particular, the present invention relates to an improved adjacent frame difference method. Background technique [0002] Intelligent video surveillance is the use of computer vision technology to analyze and judge the collected video signals without any interference, so as to realize the operations of identifying, locating, tracking and alarming the scene changes in the monitoring scene. [0003] Moving object detection is a frontier direction that has attracted much attention in the field of intelligent video surveillance, and has a very broad application prospect. At present, the methods of moving target detection mainly include adjacent frame difference method, optical flow method, background subtraction method and so on. The adjacent frame difference method is aimed at two or three adjacent frames of images, and then extracts the motion area in the image through thresholding. The optical flow method estimates the motion field based on the temp...

AI Technical Summary

Problems solved by technology

The optical flow method estimates the motion field based on the temporal and spatial gradients of the image sequence. Through this estimation, the scene detection and segmentation of the moving target is carried out. The calculation amount is large and cannot meet the real-time requirements.

Images