System and method for detecting article in video scene

Abstract

The invention relates to a system and a method for detecting an article in a video scene. The method comprises the following steps of: (a) setting an empty background model and receiving a first frame of scene in N frames of colored video scenes; (b) comparing each pixel of a second frame of scene serving as the current scene with a corresponding pixel in the background model of the first frame of scene, and judging whether a difference of the pixel values and a brightness difference value are smaller than or equal to preset threshold values; (c) if the difference of the pixel values and the brightness difference value are smaller than or equal to the preset threshold values, determining that the pixel is a background pixel and adding the pixel into the background model so as to generate a new background model, wherein all background pixels form a background article; or (d) if the difference of the pixel values and the brightness difference value are not smaller than or equal to the preset threshold values, determining that the pixel is a foreground pixel, wherein all foreground pixels form a foreground article; and (e) taking one of other frames of scenes in the N frames of scenes as the current scene, taking the background model obtained by detecting all the scenes before the current scene as the current background model, and performing the steps from (b) to (d) so as to detect the foreground article and the background article in each scene. By the invention, a dynamic background can be updated in real time.

Description

technical field [0001] The present invention relates to a detection system and method, in particular to an object detection system and method in a video scene. Background technique [0002] At present, the detection of foreground objects in the video scene obtained by shooting the monitoring area can be realized through the existing technology, but the existing technology does not have a standard test specification or process, and there is no way to overcome the following problems : Brightness changes in the monitoring area, periodic object interference, image shaking, etc. Contents of the invention [0003] In view of the above, it is necessary to provide a method for object detection in a video scene, by establishing a background model and using a foreground detection algorithm to detect foreground objects in a video scene, and tolerating changes in light to update the dynamic background , in order to achieve effective detection, monitoring and other purposes. [0004]...

AI Technical Summary

Problems solved by technology

[0002] At present, the detection of foreground objects in the video scene obtained by shooting the monitoring area can be realized through the existing technology, but the existing technology does not have a standard test specification or process, and there is no way to overcome the following problems : Brightness changes in the monitoring area, periodic object interference, image shaking, etc.

Images