Research on Mutual Information Edge Extraction of Visible Image Based on Texture Enhancement and Refinement
When using traditional methods to extract image edges,problems such as inaccurate image mutual information matching,low F1 value,and unsatisfactory edge detection performance may arise.A visible im-age mutual information edge extraction based on texture enhancement and refinement is proposed.The im-age is registered using mutual information registration formulas and the image is preprocessed on this ba-sis.The processing steps include image texture enhancement,image binarization,image smoothing and re-finement,and finally obtain clear data without interference.Then the two-dimensional smoothing function and Lipschitz method are used to perform edges detection and extraction on the the optimized image.The experimental results show that the F1 value of the proposed method is always higher than 0.85,with a true case rate is more than 95％ and less than 20％ in the three groups of experimental datasets.The experi-mental data indicates that the proposed method has better application performance.
国家科技期刊平台
NSTL
万方数据
