An automatic tracking and track modeling method for short-track speed skating athletes in videos

Abstract

The invention discloses an automatic tracking and track modeling method for short-track speed skating athletes in videos. The method comprises the steps of using a high-speed camera for collecting athlete training or competition videos from a boundary line overlook view angle; sampling the image in the video according to a specific frequency to obtain a group of single-frame images; carrying out target detection on the athletes in the single-frame image, and determining the positions of the athletes by using a square frame; matching the target detection results belonging to the same athlete among the frames to obtain a tracking result belonging to the same athlete target; mapping the tracking result of each athlete to the ice surface, carrying out curve fitting, and drawing a smooth motiontrack; estimating and outputting the motion parameters of the athlete according to the track equation. The short-track speed-skating game analysis system has no interference on games, can directly collect and analyze the positions of athletes, greatly improves the automation level, reduces the manpower consumption, can quickly feed back related kinematics parameters after games, improves the speed, and provides possibility for real-time short-track speed-skating game analysis.

Description

technical field [0001] The invention relates to the technical field of motion image analysis, in particular to an automatic tracking and trajectory modeling method for short-track speed skaters in videos. Background technique [0002] Existing motion image analysis technology uses wearable devices as the information source, and tracks the devices through wireless signals such as wifi or bluetooth to obtain the athlete's position. The main disadvantage is that the wearable device needs to be embedded in the athlete's equipment, which is not allowed in the general game scene. The non-contact method is mainly based on motion image analysis, but the positioning accuracy of the traditional method is not high enough. Contents of the invention [0003] The purpose of the present invention is to provide a method for automatic tracking and trajectory modeling of short-track speed skaters in the video. Compared with the prior art, it does not interfere with the athletes' competitio...

AI Technical Summary

Problems solved by technology

The main disadvantage is that the wearable device needs to be embedded in the athlete's equipment, which is not allowed in the general game scene

Non-contact methods are mainly based on motion image analysis, but the positioning accuracy of traditional methods is not high enough

Images