Strip steel surface defect image mosaic method based on area CCD

The strip steel surface defects which collected by area CCD are not continuous image, there are overlap between each other. In order to join the overlapped images properly, and meantime maintain the original image feature, a weighted average fusion method based on provincial characteristics is adapted. A linear interpolation image fusion method is proposed for improving the contrast of the fused image, enhance the high frequency information, and eliminate matching trace. With linear interpolation image fusion, the pixels gray value of overlap area is replaced by the corresponding weighted average of gray value in two images. Experimental results show that the fusion method can meet the requirements of strip steel defect detection. And it can effectively eliminate the traces of image mosaic, make the image clearer. This made a foundation for the strip surface defect identification and quality evaluation.