Camera scene change detecting method based on image angular point set characteristic

Abstract

The invention discloses a camera scene change detecting method based on the image angular point set characteristic. The camera scene change detecting method based on the image angular point set characteristic comprises the following steps that firstly, image angular point set information is obtained; secondly, an angular point set matching template is trained; thirdly, whether scene change occurs is judged. According to the camera scene change detecting method based on the image angular point set characteristic, the image angular point set information characteristic and the mathematical statistic probability template are combined to be used for judging whether scene change occurs within the monitoring range of a camera, and the image angular point set information is adopted ruing processing; compared with an existing method, only angular point information is adopted, the interference of pseudo angular points caused by accidental factors can be reduced, and the subsequent angular point matching is facilitated due to the fact that the angular point set information has the statistical probability stability; the overall angular point set probability matching template which is trained in advance is essentially manufactured according to statistical probability events in an angular point appearing area; in this way, the influence of the external accidental factors can be eliminated to a great extent, the method can be adapted to the practical engineering application environment, and the accurate detection effect can be achieved.

Description

technical field [0001] The invention belongs to the technical field of image processing, specifically a camera scene change detection method based on image corner set features, which is mainly used for real-time monitoring of the scene change in video surveillance, thereby judging whether the monitoring state of the surveillance camera is meet the requirements. Background technique [0002] With the wide application of video surveillance technology, automatic detection of abnormal events based on image processing has become a research hotspot. At present, most automatic video detection technologies detect abnormal events based on specific detection areas. Once the monitoring range of the camera changes or the scene changes, the detection area will inevitably be misplaced, thereby affecting the results of automatic event detection. Therefore, it is of great social significance and application prospect to study how to timely and accurately detect the scene change event of the...

AI Technical Summary

Problems solved by technology

[0006] The scene change detection methods mentioned in the existing literature are purely for the extraction and matching of image features in an ideal environment, without considering the interference of external factors and the influence of noise, and the detection effect is not good when directly applied in the real environment , especially in environments such as highways with many interference factors and poor environmental conditions, it is an urgent problem to achieve more accurate scene change detection.

Images