Target relay tracking algorithm based on intelligent street lamp partners

Abstract

The present invention discloses a target relay tracking algorithm based on intelligent street lamp partners. According to the present invention, a target is detected by utilizing a deep network model, and secondly, the features of the target are extracted by utilizing another one deep model. An intelligent street lamp partner can snapshot the images of a plurality of visual angles of an object, and save the features extracted at each visual angle. A system detects the target at set intervals, and fuses a detection result and a tracking result. When the object moves out from the visual field of the current intelligent street lamp partner and enters in the visual field of the next intelligent street lamp partner, the next intelligent street lamp partner similarly detects the target and extracts the features, and the extracted features are compared with the features saved by the last intelligent street lamp partner, so that whether the detected objects are the same object can be determined, and the target relay tracking is realized.

Description

technical field [0001] The invention relates to a target relay tracking algorithm of a smart street lamp companion, which belongs to the technical field of computer vision. Background technique [0002] Video surveillance is one of the important means to maintain public security, and target tracking is a basic research topic for video processing. In a complex environment, the problems faced by traditional tracking algorithms are: 1. Using a single camera for tracking, the field of view is limited. 2. Similar targets cause some interference to the tracking process. 3. The target is blocked or deformed multiple times during the movement. 4. Low image resolution or low image quality will affect subsequent processing. Multiple cameras can track the target in a wide area, and the obtained target trajectory is more complete and more practical; but it also brings some new problems, such as data fusion, target matching, etc., which are currently being studied However, this does ...

AI Technical Summary

Problems solved by technology

In a complex environment, the problems faced by traditional tracking algorithms are: 1. Using a single camera for tracking, the field of view is limited

2. Similar targets cause certain interference to the tracking process

4. Low image resolution or low image quality affects subsequent processing

Multiple cameras can track the target in a wide area, and the obtained target trajectory is more complete and more practical; but it also brings some new problems, such as data fusion, target matching, etc., which are currently being studied However, this does not prevent the multi-camera system from gradually becoming the mainstream trend in the field of monitoring and tracking.

Images