| Summary: | 碩士 === 淡江大學 === 航空太空工程學系 === 90 === Grey theory has been developed for years and has successfully applied to various fields such as economy, society,engineering, and statistics, etc. The image-processing algorithm used in this research is based on the concept of gray theory. An N-power type of gray relational edge detection model was proposed and applied in image processing. The transform formulation used in this model is a type of inverse function and the principle of detecting edge is similar to gradient operator. However, the coefficients of the mask vary with positions.
Algorithm used for edge detection and noise removing has been proposed. Two edge detection methods, including gray relational series edge detection and gray relational array edge detection methods were detailed in the thesis. The method used in noise removing, is similar to the median filter in that they both take the median value of a working mask.
According to the results, the capability of edge detection of this research is better than that of Laplace filter and Prewitt filter, but close to Sobel filter. But the behavior of characters and texture is better than Sobel filter. Removed noise method is similar to median filter. They both take median value of a mask.
|
|---|
