The Research Of Denoising And Edge Detection Method In Image Measurement

During image measurement process, the acquired image is susceptible to noise which brings difficulty to the follow-up image feature extraction and analysis. Furthermore, the performance of edge detection methods will also affect the accuracy of the image feature extraction directly. So how to de-noise and edge extraction effectively is of great significance in image measurement. This paper focuses on the research and improvement of denoising and edge detection methods in image measurement.First, do a study of the traditional denoising methods especially of wavelet threshold denoising method, and then do a comparison and analysis of their denoising effects. Aim at the shortage of the traditional wavelet thresholding denoising methods, this paper proposes an improved wavelet thresholding method which combines high-density wavelet transform with bivariate shrinkage function. Via experiment verification and comparison, the improved method can achieve higher Peak Signal to Noise Ratio, has better denoising effect.Then, do a comparative analysis of several typical edge detection methods. Do a study of the principle and process of wavelet edge detection. Then presents an improved wavelet edge detection method based on a'trous algorithm. After the experiment simulation, the improved edge detection method based on wavelet transform can extract the edge features effectively, keep the image details and have a good continuity.At last, do a research of the hardware structure of the image measurement system and analyzes the main function of the hardware modules and its impacts. Through the parameters measurement of micro-gear, get the main dimension parameters. Then compare them with the results obtained under other typical methods, they have higher accuracy and the minimum relative error, further verified the feasibility and...