Corner Detection of Line Fitting Based on Optimization of Parallelogram Diagonal Theory
In the process of Chinese character image recognition,corner detection helps to analyze the characteristics of Chi-nese characters and reflect the important information of each part of the Chinese character image.In the previous corner detection al-gorithms,the algorithm for judging corners by curvature has good results,but the detection of corners at obtuse corners or arcs is not accurate enough.On the basis of self-adaptive straight line fitting corner detection,the final corner set is filtered out by estimating the diagonal curvature of the parallelogram at the corners of the arc.By comparing the existing corner detection algorithms,on the basis of taking into account the time efficiency,the proposed method has higher accuracy and better application effect for the corner detection of Chinese characters.
万方数据
维普
