The Improved Threshold Function Combined With Morphology Edge Detection Method

When considering digital image processing, noise filtering is a kind ofoperation that not be ignore. Generally produces in the process of imagetransmission, storage and other pulse interference, as the interference intensity isdifferent, the ultimate digital image damage is also different. Edge detectionalgorithms in the process of implementation will take into account the smallconcentration of noise interference. Before edge detection, need for imagede-noising. Make the image noise smaller, conducive to edge detection.Wavelet threshold de-noising method is easy to implement, the operation issimple and has more extensive adaptability. But they also have defects. Hardthreshold function is discontinuous, between soft threshold function and real waveletcoefficients of image wavelet coefficients are biased. This paper constructs a new atthe origin smooth threshold function fully, have better continuity, almost everywhereis smooth, and this function is simple. A new method of image de-noising achievedsatisfactory effect.The edge detection algorithms of Mathematical morphology does not based ondifferential operation as a result of the detection principle, so it has a good effect onnoise resistance. As a new type of nonlinear processing method, mathematicalmorphology operation and analysis mathematical set for processing images. Usingthe traditional morphology algorithm for image edge detection, if the noise densityis higher, the effect is not ideal. So before the image edge detection using theimproved wavelet threshold de-noising method to deal with the noise of image,obtain after de-noising image, finally, do the image edge detection with morphology.Experimental results better than only use morphological edge detection results,proves the validity of this method.