Study For Application In STM Image Of The Denoising Method Based On Wavelet Transform

College of Mathematics and Physics in Chongqing University researched and developed STM.IPC-205B. In the process of acquiring or transferring, due to the images maybe suffer some interference which come from environmental noise or equipment noise, it is not easy to recruit the real characteristic of the original signal, and will affect the analysis and disposal of image signals. Because STM is different from SEM (Scanning electronic microscope) which has the expert gallery, judging the STM images entirely relies on the experience or subjective opinion of experimenter, which can be caused many errors. So it is necessary to make further improvement in image processing technology.Nowadays, the image enhancement of most commercial STM is processed by means of off-line handwork. It usually enhances STM image by using all kinds of filters in spatial domain and frequency domain combining with Photoshop. But this method can not manage the scanning images timely. So, an improved image enhancement method in wavelet image based on spatial domain correlation is proposed in this paper.In this paper, the leading work include: According to the characteristics of STM and local analyzing in space-frequency of wavelet transform, an algorithm of denoising method for STM image is proposed. We put forward a real-time improved image enhancement method based on multi-scale correlation in wavelet domain based the chara-cteristic of STM image noise in this paper. First, calculate the histogram and cumulative histogram of the wavelet multi-scale correlation. And then, determine the signal-noise threshold in each band by analyzing noise intensity of the cumulative histogram. At last, use an improved soft-thresholding function to threshold the wavelet coefficients of signal. Experiments show that this method can remove noise and preserve significant details, and it has a well execution speed. All of this makes it possible to process STM image timely and on-line which can reduce jamming influence and enhance STM image quality. It affords a good base for latter analyzing and explaining images.