GM(1,1) model edge detection method based on accumulation method

Edge detection is a crucial step in image processing. The accuracy and clarity of edge positioning directly affect the accuracy of subsequent processing. The original GM (1,1) prediction model is morbidity, the parameter estimation of the accumulation method can be used to instead of the parameter estimation of the least squares method, the original data can be added with the offset and multiply for transformation to solve it. In this paper, an edge detection algorithm for GM(1,1) prediction model based on accumulative method is proposed. First, Pre-processing the original grayscale data of the image, then the appropriate original template is selected. The intensional prediction formula of the GM(1,1) model of the accumulation method is used to obtains the prediction image, divides the prediction image into positive and negative subgraphs and performs double threshold segmentation, and then combines the divided positive and negative subgraphs to obtain image edges. Finally, this paper analyzes the effects of different original templates and compares them with the original GM (1,1) prediction model edge detection algorithm. It proves that the proposed algorithm not only eliminates the morbidity of the model but also obtains clearer and more complete edges, and the speed is also faster.