Object tracking method and apparatus

Abstract

A method and device for tracking an object using depth images or parallax images are provided. The method includes: for a newly detected first object, using a depth image sequence including the first object, extracting short-term features and long-term features of the first object, and storing them as a short-term feature template and a long-term feature of the first object. Feature template; during the tracking process of the first object, for each frame of the image to be analyzed, tracking is performed based on the match between the extracted short-term features of the candidate object and the stored short-term feature template of the first object; if based on the first object The short-term feature template matching fails, and the first object is identified as the missing first object; for the subsequently newly detected second object, the short-term features and long-term features of the second object are extracted; and based on the second object Matching between the short-term feature and the long-term feature and the short-term feature template and the long-term feature template of the first object identified as missing determines whether the second object is the missing first object.

Description

technical field [0001] The present invention generally relates to object tracking methods and devices, and more particularly to methods and devices for tracking one or more objects using depth images. Background technique [0002] Object tracking technology such as human tracking technology is a key technology required for applications such as human positioning services, video surveillance, robot navigation, and intelligent driving assistance systems. At present, it is still a challenging technical topic, mainly because the posture and appearance of people in practical applications vary greatly, and occlusion problems often occur. In real scenes, occlusion problems occur from time to time, especially when people are in motion, or other people or objects are in motion. In indoor scenes, due to the complex background and numerous people, occlusion occurs more frequently, which often leads to the failure of tracking. [0003] Patent document US88073197B2 proposes a method for...

AI Technical Summary

Problems solved by technology

The method based on the horizontal plane view proposed in this non-patent literature (3-D to 2-D coordinate transformation, histogram statistics, that is, calculating features on the plane view) can alleviate the influence of occlusion to a certain extent, and it can be used in living rooms, lobbies, etc. Such scenes with relatively simple background, but for complex office scenes with many partitions, for example, the problem of tracking failure will occur

Images