Detection of internal mechanical cracks in corn seeds based on data fusion technology

In order to further study the mechanism and detection technology of internal cracks during threshing process and transportation of corn seeds, an edge detection method with data fusion based on stereomicroscope was proposed. The image edges of corn seeds with mechanical damage were respectively detected by mathematical morphology and Sobel, and fusion rules were set up accordingly. The two results from above methods were then processed by fusion based on wavelet transform to generation a new image. The feature information of inner mechanical damage from new image of corn seeds was extracted. Results showed that the proposed method had the advantages of two edge detection methods, which could improve the accuracy of edge detection and reduce noises while accurately extracting internal mechanical traits of corn seeds. The new method could obtain better effect than single traditional edge detection method.