Study of image fusion algorithm to edge erosion of titanium alloy with DR detection

This paper studies the image fusion algorithm of edge corrosion on large thickness difference titanium alloy with DR detection. First collected the DR digital image of large thickness difference workpiece, then calculated the edge width of image with ImageJ, and finally handled with edge corrosion area through the image fusion technology of PCNN. The information of fused image was more wealth than the original image so that could reduce the width of edge erosion, which could restore covered defects and achieve the aim of improving the detection rate of defects. This paper is based on the comparison between PCNN algorithm and another fusion algorithms, and assessed the quality of fused image with the width of edge erosion, sensitivity of thickness, resolution of IQI, information entropy(ENTROPHY), mutual information(MI), correlation coefficient(CL). The result of research shows that PCNN algorithm is superior to another algorithms in the aspect of information entropy, mutual information, which mainly embodies that PCNN image fusion algorithm can get more detailed information of the original image. Besides, PCNN algorithm can effectively reduce image´s edge erosion, improves image resolution and image thickness sensitivity, which can restore more overwritten defects with the aim to improve the defect detection rate.