Focus target detection method based on air-ground collaborative view angle

Abstract

The invention discloses a focus target detection method based on an air-ground cooperative visual angle, and the method comprises the steps: carrying out the detection of a focus target based on a first visual angle of a high-altitude camera carried by a high-altitude visual angle unmanned plane and a first visual angle of a wearing camera worn by a plurality of ground pedestrians through the air-ground cooperative visual angle; positioning of a focus target is realized by detecting, associating and fusing visual attention scenes of multiple persons in the scene; the method comprises the following steps: S1, inputting N + 1 videos with time synchronization, wherein the duration is T; S2, segmenting all the videos obtained in the step S1 into short segments with the duration of T by using a sliding window technology; and S3, obtaining a candidate focus target from each frame of the current fragment by using a human object detection algorithm and a multi-view data fusion method, and the like. According to the invention, surrounding personnel and environmental details at the same position are captured through a plurality of cameras of a first view angle, and detection of the focus target in the scene is realized by using the air-ground collaborative view angle.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a focus target detection method based on an empty collaborative perspective. Background technique [0002] The background technology involved in the present invention includes: [0003] (1) Video joint segmentation and saliency detection: Video joint segmentation is to simultaneously segment objects of a common category from two or more videos. In practice, almost all existing video collaborative segmentation methods are based on the appearance information of objects. Some methods also utilize supplementary motion features and consider temporal information for joint segmentation of objects (ref. ). Unlike this method, video co-segmentation looks for all common objects, including human subjects, in different videos, without Consider whether the detected subject has attracted the interest of other surrounding subjects. [0004] (2) Video saliency detection: Video salie...

AI Technical Summary

Problems solved by technology

[0006] (4) Multi-view video analysis: The collaborative analysis of multi-view videos is of great significance. However, the existing multi-view video collaborative analysis has a limited field of view, and the need to wear wearable cameras affects its usability in video surveillance systems.

Images