Sports Training Aid Using Image Processing

使用图像处理的运动训练辅助工具

• 批准号: 11680209
• 负责人: NISHIHARA Akinori
• 金额: $ 2.18万
• 依托单位: TOKYO INSTITUTE OF TECHNOLOGY
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 1999
• 资助国家: 日本
• 起止时间: 1999 至 2000
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-11680209/
• 关键词: Image Processing; Sports Training; Detection of Player's Position; Badminton; Trace; DSP

Two systems to detect the position of player in a badminton or tennis court are developed and player movement is analyzed.The first system can detect player position in real time using an interface using a DSP.Position information is transferred to a personal computer. First an image taken by a camera which is set above the court is subtracted by an image of the court without any players. The player position is detected using a suitable filter. At first indoor games such as badminton is the main target, but the system is upgraded by introducing a background updating algorithm so that it can be used for outdoor games like tennis where the court brightness varies depending on the weather. A noise reduction filter and an algorithm to discard flames which seem to be erroneous are also introduced. The system can process 7 to 8 flames/sec of 128×128 pixel images. Training can be improved by feedback of position and movement information of the player.The second system handles higher resolutio … More n of 360×240 at the expense of real-time property, and all the processing is performed by software on a PC.In this case each pixel in the difference image is quantized to binary number and labeled so as to remove noise component. Square area around the center of gravity of the player image is taken as an initial template, which is updated in the following flames by the template matching.Using these systems traces of beginners and experts in badminton singles and doubles. It is charified that, in the singles experts move faster in forward-backward direction and no significant difference in left-right movement, while in the doubles experts move faster in left-right direction and no significant difference in forward-backward movement. Average position also differs between beginners and experts. These results suggest the important points for beginners to practice.Blood lactate is shown to have strong influence on total distance and mean velocity, which suggests the possibility that the trace analysis can estimate the blood lactate. Less

开发了两个系统来检测羽毛球场或网球场中运动员的位置，并分析运动员的运动。第一个系统可以使用 DSP 接口实时检测运动员位置。位置信息传输到个人计算机。首先，将设置在球场上方的摄像机拍摄的图像减去没有任何球员的球场图像。使用合适的过滤器检测玩家位置。最初以羽毛球等室内比赛为主要目标，但系统通过引入后台更新算法进行升级，以便可以用于网球等球场亮度随天气变化的室外比赛。还介绍了降噪滤波器和丢弃似乎错误的火焰的算法。该系统可以每秒处理 7 至 8 个火焰的 128×128 像素图像。通过反馈玩家的位置和运动信息可以改进训练。第二个系统以牺牲实时性为代价处理更高分辨率的360×240，所有处理都是由PC上的软件执行。在这种情况下，差异图像中的每个像素都被量化为二进制数并进行标记，以消除噪声分量。以运动员图像重心周围的正方形区域作为初始模板，通过模板匹配在后续比赛中更新该模板。利用这些系统跟踪羽毛球单打和双打的初学者和专家的痕迹。据了解，单打选手前后移动速度较快，左右移动无显着差异；双打选手左右移动速度较快，前后移动无显着差异。初学者和专家之间的平均位置也不同。这些结果表明了初学者练习的要点。血乳酸被证明对总距离和平均速度有很大影响，这表明痕量分析可以估计血乳酸的可能性。较少的