Research on Multi-level Segmentation Method of Digital Image Based on Minimum Spanning Tree
Nowadays,due to the large segmentation error caused by the image discretization,the research of the multi-level segmentation method of digital image based on the minimum genera-ting tree is proposed.First of all,based on the minimum generation tree for digital image pre-processing,the overall image of feature extraction,secondly,build the image multi-level seg-mentation standard,based on different scene requirements,to define the weights,to ensure that the final segmentation results within a high degree of similarity,finally,realize the digital image multi-level segmentation.In order to verify the effectiveness of the method in the paper,a comparative test is needed.The experimental results:the number of digital image based on the minimum spanning tree is more in line with human visual characteristics,the segmentation effect is more ideal,and has good performance.
NETL
NSTL
万方数据
