Based On Fuzzy Inference Of The Noise Image Edge Detection Techniques

Edge detection is one of the most fundamentals in image processing and analyzing, which is still unsolved perfectly. Image's edges include image's features such as position and outline, which belong to the fundamental features. Edge detection is widely used in image analysis and processing such as feature description, image segmentation, image enhancement, pattern recognition and image compression etc., so edge detection is the research hotspot in the technology of image processing and analysis all the while, on which the new theory and methods are put forward constantly.Because of the influence of the external various factors, image is subjected the noise of interference easily in the process of acquisition and transmission. The edge frequency is similar to the noise, so some problems exist in this process, such as the false edge, missing edge and the edge is just not the single pixel edge. Therefore people pay more and more attentions to the technology of edge detection in noisy image.Some edge detection algorithms are researched, including Roberts, Sobel, Prewitt, LOG and Canny operators, and algorithms which gained more attention, such as morphology, wavelet transform, neural network and fuzzy reasoning. Experiments are taken to the traditional operator in this paper. Roberts operator is very sensitive to noise. Noise inhibiting ability of Sobel and Prewitt operators is better than Roberts operator, because these operators use weighted average before differential. In order to inhibit noise, the operators of LOG and Canny smooth the noisy image and some edge is missing in this processing.In image processing and analysis field, because of the fuzziness that the image edge itself has, obvious superiority is showed by applying the fuzzy theory to edge detection of image. Therefore, the general theory of fuzzy logic is introduced in this paper, based on which, the relative algorithms about edge detection with fuzzy theory are made a research. Tao use the fuzzy theory to do the detection, which can separate the weak border image from the background much better. But some problems exist in this process, such as the noise inhibiting ability is poor and the edge is just not the single pixel edge. It's defective that Tao use the single gradient feature of the pixel to express the information of edge. Also the rules of this algorithm are complex, which lead more calculation. To detect edge in noisy images, the related gradient is proposed to distinguish noise from edge in this paper, which is combined with the basic gradient of Tao's algorithm. A method of edge detection based on fuzzy reasoning is proposed in this paper. The basic gradient and the related gradients are considered as the input variables of the fuzzy reasoning system, meanwhile fuzzy rules are executed so as to get a noisy membership degree in order to remove the false edge. A processing of thinning the result is proposed to acquire the better result. The fuzzy reasoning in this paper analyses the difference between edge and noise, so the result of experiment is better.The experimental results, which is acquired by Matlab 7.0, show that compared with other edge detection operators, more edge information is acquired and more noise is suppressed by the proposed method.