An Edge Detection Algorithm Based on Adaptive Threshold

The threshold need artificial hypothesis in the existing edge detection algorithm. In response to this phenomenon, we propose an edge detection algorithm is based on adaptive threshold. Article algorithm was studying the characteristics of wavelet transform. And it was combined with the traditional edge detection algorithm to reset the threshold. So we can use this adaptive Threshold to search the edge signal. Experimental results show that the results consistent with the expected results. Article algorithm can kept a better edge detection effect under different condition. And it cans effective extract the edges of each frequency sub-band. Introduction The edge is one of the most fundamental features of image. It is the junction of the property area to another property area. It is the place where regional attribute mutations. So it has a wealth of information. Therefore, Edge image extraction having a key role in computer processing. Nowadays, have a lot of edge extraction operators [1-3]. For example: Roberts, Sobel, Canny, etc. It strengthened the processing of the image to a certain extent. But the disadvantage is susceptible to noise interference. In recent years, researchers proposed many new algorithms[4,5]. For example: Edge detection algorithm base on wavelet Transform, Edge detection algorithm base on mathematical morphology, Edge detection algorithm base on energy guidelines, etc. These methods have made better results than traditional methods. However, there exists a great mount of computations. The computing speed is not acceptable for practice, either. Focusing on the problem, we have research of the traditional edge detection algorithm and wavelet transform[6,7].And proposed an edge detection algorithm based on adaptive threshold. It was redefines the computing and threshold of the image edge. High frequency sub-bands for different directions using different search window to get the wavelet coefficients. And adaptive adjustment of global threshold and local threshold to enhance the effect of edge detection. At the same time, In order to avoid the maximum positioning error in the threshold of the process. Article algorithm was according to the characteristics of edge gradient and gradient variance to improve threshold. Traditional edge detection algorithm based on wavelet By wavelet transform, we can get that dt a b t t f a C b a W       ) ( ) ( | | 1 1 ) , ( f 