基于红外光谱图像的运动损伤自动检测研究

Research on automatic detection of sports injury based on infrared spectrum image

为能够提升运动损伤自动检测准确性,节省检测时间,提出基于红外光谱图像的运动损伤自动检测。利用红外光谱图像采集系统采集运动损伤红外光谱图像;利用小波变换去噪处理红外光谱图像;通过结合单阈值增强算法、双阈值增强算法与自适应增强算法,增强去噪后的红外光谱图像;通过计算增强后红外光谱图像中运动损伤区域可靠性,确定运动损伤区域;利用小波分析法提取确定区域内运动损伤信号特征;构建自适应神经网络模型,以运动损伤信号特征为输入量,输出结果即运动损伤自动检测结果。实验证明:所研究方法能够有效检测出运动损伤区域,运动损伤区域确定误差率低;运动损伤自动检测结果准确性高达98%,自动检测速度快;加入不同程度噪声后,受噪声影响较小,抗干扰能力强。

The research of sports injury automatic detection based on infrared spectrum image can improve the accuracy of sports injury automatic detection and save detection time. The infrared spectrum image of sports injury is collected by infrared spectrum image acquisition system;the infrared spectrum image is denoised by wavelet transform;the denoised infrared spectrum image is enhanced by combining single threshold enhancement algorithm,double threshold enhancement algorithm and adaptive enhancement algorithm;the motion injury area is determined by calculating the reliability of the motion damage area in the enhanced infrared spectrum image;The wavelet analysis method is used to extract and determine the characteristics of the sports injury signal in the region,and the adaptive neural network model is constructed. The feature of the sports injury signal is taken as the input,and the output result is the automatic detection result of the sports injury. The experimental results show that: the method can effectively detect the sports injury area,and the error rate of the determination of the sports injury area is low;the accuracy of the sports injury automatic detection is as high as 98%,and the automatic detection speed is fast;after adding different levels of noise,it is less affected by the noise and has strong anti-interference ability.