基于函数型数据时间序列建模的齿轮破损图像识别

当前齿轮破损图像识别未拟合周期性动作捕捉数据,导致齿轮破损图像的识别效率较低,图像边缘轮廓中细节信息丢失,造成识别误差较大、识别效果较差.提出基于函数型数据时间序列建模的齿轮破损图像识别方法,通过小波变换方法提取齿轮破损图像中存在的低频成分,做双直方图均匀化处理,采用高斯高通滤波器提取齿轮破损图像中存在的高频成分,增强...

Image Recognition of Gear Damage Based on Function Data Time Series Modeling

The current gear damage image recognition does not fit the periodic action capture data,resulting in the low efficiency of gear damage image recognition and the loss of detail information in image edge contour,resulting in large recognition error and poor recognition effect.A method of gear damage image recognition based on function data time series modeling is proposed.Wavelet transform is used to extract the low-frequency components in the damaged gear image,and the double histogram homogenization is performed.The high-frequency components in the damaged gear image are extracted by Gaussian high-pass filter to enhance the detailed information in the damaged gear image.Function data analysis method is used to fit the action capture data.According to the obtained data,the period time series is constructed,and the hidden Markov model is obtained.Through the maximum likelihood estimation function,the matching degree between the test model and the sample is calculated to realize the recognition of gear damage image.The experimental results show that the recognition error of the proposed method is small and the recognition effect is good,which can effectively improve the recognition efficiency of gear damage image.