The detection method and device of a colorless scene video on specific target areas

Abstract

The present invention provides a method and device for detecting a specific target area in a colorless scene video. The method and device comprehensively utilize different spatiotemporal information such as brightness, texture, outline and dynamics in a colorless video to construct a video scene area type Descriptor, in order to detect and determine the specific target area, the feature extraction and calculation method on each information channel of the video scene area type descriptor is simple and efficient, with good real-time performance, high learning efficiency of video samples, and real-time online training; in addition, when the method and device calculate the type value of each channel of the video scene area type descriptor, it also introduces a parallel computing method, which significantly improves the overall computing efficiency of the method, thereby overcoming the existing method. In the color video, when detecting a specific target area, the calculation efficiency is low, the real-time performance is poor, and even the problem of failure.

Description

technical field [0001] The present invention relates to the technical field of video processing, in particular to a method and device for detecting a specific target area in a colorless scene video. Background technique [0002] With the vigorous development of visual mobile robots, unmanned vehicles, unmanned aerial vehicles and other technologies, the importance of vision-based scene analysis technology has become increasingly prominent. The detection and identification of specific target areas is a key technology in vision-based scene analysis, because the quality of detection and identification of specific target areas (such as common obstacles) will directly affect mobile robots, unmanned vehicles, and unmanned vehicles. The operation mode and operation safety of human-machine and other equipment. [0003] Since unmanned vehicles, unmanned aerial vehicles and other equipment usually run at high speeds, the analysis and processing efficiency of real-time video data gene...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the defects existing in the prior art, the present invention provides a method and device for detecting a specific target area in a colorless scene video. Spatio-temporal information constructs a video scene area type descriptor to detect and determine a specific target area. The feature extraction and calculation methods on each information channel of the video scene area type descriptor are simple, efficient, and real-time. The learning efficiency of video samples high, and can realize real-time online training; in addition, when the method and device calculate the type value of each channel of the video scene area type descriptor, it also introduces a parallel computing method, which significantly improves the overall computing efficiency of the method, thus overcoming Solve the problems of low computational efficiency, poor real-time performance, and even failure when the existing method is implemented to detect a specific target area in a colorless video

Images