CAJ | 학술논문

在复杂运动训练场背景下对运动员的肌体行为特征的准确检测是实现运动技术行为跟踪和识别的重要基础，在比赛监测和训练指导方法具有重要的应用价值。运动场景具有不确定性，图像背景具有跳变性，检测较为困难。提出一种基于灰色关联的运动训练场景背景下的运动员肌体行为特征检测算法，人体在进行复杂场景模式的训练和肌体作用下，特别是在如足球运动、跑步运动时，骨骼围绕着关节作为支撑进行矢状面内的运动，通过对人体肌体的驱动模式动力学进行建模分析，采用先分割再填充的方法，获得不重影、不抖动的图像，通过灰色关联分析准确识别运动训练背景下的人体肌体行为特征，编制运动员的肌体行为识别软件系统，通过运动录像解析系统准确及时地反馈运动信息，提高体育运动训练的科学化水平，促进训练和比赛质量的提高。实验结果表明，该算法能满足运动场景中的全方位多尺度信息采集的肌体动作识别的要求，误检率较传统算法低，性能较好。
재복잡운동훈련장배경하대운동원적기체행위특정적준학검측시실현운동기술행위근종화식별적중요기출，재비새감측화훈련지도방법구유중요적응용개치。운동장경구유불학정성，도상배경구유도변성，검측교위곤난。제출일충기우회색관련적운동훈련장경배경하적운동원기체행위특정검측산법，인체재진행복잡장경모식적훈련화기체작용하，특별시재여족구운동、포보운동시，골격위요착관절작위지탱진행시상면내적운동，통과대인체기체적구동모식동역학진행건모분석，채용선분할재전충적방법，획득불중영、불두동적도상，통과회색관련분석준학식별운동훈련배경하적인체기체행위특정，편제운동원적기체행위식별연건계통，통과운동록상해석계통준학급시지반궤운동신식，제고체육운동훈련적과학화수평，촉진훈련화비새질량적제고。실험결과표명，해산법능만족운동장경중적전방위다척도신식채집적기체동작식별적요구，오검솔교전통산법저，성능교호。
Accurate detection of body behavior characteristic of athletes in sports training field is an important basis of mo?tion tracking and recognition, it has important application value in the competition and training methods monitoring. The sports scene is uncertain, and the image background has jumped degeneration, so the detection is difficult. A body behavior characteristics detection algorithm is proposed based on grey correlation in sports training background, especially in sports such as football, running, the segmentation of image is obtained after filling. The image without ghosting and jitter is collect?ed. The body behavior detection is taken based on grey relational analysis, and body behavior recognition software system is designed. The sports video analysis system can accurately and timely feedback motion information. The scientific sports training level are improved. It can promote the quality of training and the competition to improve. The experimental results show that algorithm can meet the full range of motion in the scene of multi-scale information collection, the recognition and detection performance is good. The false detection rate i8s lower than the traditional algorithm, and the performance is bet?ter.