Research of high-speed railway image noise removal based on fast recursive GCV threshold function

This paper introduces the wavelet threshold de-noising process and the principles of traditional GCV threshold to solve the problems about the distortion and noise removal of the images collected in the high-speed train to monitor real-time for preventing foreign body from the track. It puts forward an image de-noising method on fast recursive GCV threshold function after introduced integer wavelet transform and the thought of merging recursive operations in this paper. The de-noising effects was analyzed and compared between tradition threshold and FR-GCV threshold function by simulation and experiments, it turns out that FR-GCV threshold function can not only retain the characteristic of obvious de-noising effects, but also avoid fairly well to use the complex algorithm. It has a certain guiding significance in practice.