Studies On Characteristic And Identification Of Internal Mechanical Cracks Of Corn Seeds

Internal damage of corn seeds that is internal mechanical cracks is produced by mechanical threshing, it not only make corn seeds compressive strength decreased, processing process easily broken, and storage process easily moisture absorption, mildew and produce pests, so seeds germinate and the activity are influenced, but also is easy to find by naked eye and has more potential hazards. Research contents attach to the National Natural Science Fund (50675143) and Doctoral Fund Project of the Ministry of Education (200801570007). Internal mechanical crack of corn seeds as the research object, based on the study of bud experiment, producing, expanding the mechanism, the extraction and recognition of mechanical cracks were studied by using image processing technology in corns seeds.The main research contents and conclusion are as follows:(1) Through germination experiment of corn seeds with cracks, the results showed that the influence of turtle crack on the seed germination and development was the most significant, at the same time weak germination and no sprout of corn seeds were dissected to observe, further study influence mechanism of mechanical crack on germination and development in corn seeds.(2) The morphology, creative position and propagation direction of the internal stress cracks in corn seeds were observed through stereomicroscope, the results showed that mechanical cracks of corn seeds mainly concentrated in crown and back, and most cracks only passed through the endosperm, less injuryed embryos.(3) Producing mechanism of inner mechanical cracks of corn seeds was researched by means of cottrell piling up of dislocation theory, and the formula of expanding speed and propagation route dimension of inner mechanical cracks were inferred based on theory of griffith energy balance and fractal geometry, moreover, establish4kinds of brittle fracture extension model of mechanical cracks in corn seeds and compute their fractal dimension.(4) According to the process of image testing, to solve the blur and noise problem crack images appearing in corn seeds, new images achieved by wavelet data fusion method had higher PSNR and RMS, less noise.(5) Based on the lower precision of image edge detection algorithm, new edge detection method of couple fractional order differential theory and traditional first order different edge detection operators were introduced. Accuracy rate of the edge character by model extraction was higher and had certain inhibition for noise, besides, made up some shortage of edge detection method existing.(6) In order to solve mutual stick together of corn seed image, active active contour model of the new image segmentation algorithm was introduced, segmentation of accuracy rate is91.5%, and boundaries had not obvious deformation after partition.(7) This paper put forward detection method based on data fusion internal mechanical crackof corn seeds. Results showed that RMS, entropy, PSNR and edge keeping degree of the edge detection based on data fusion achieved optimal effect, and image edge was clear, continuous and less noise after merging.(8) Considering white impact area of corn seeds crown and fractal dimension of cracks as characters, internal mechanical crack of corn seeds implemented identification research, and identification accuracy of the two methods was higher; white impact area of corn seeds and fractal dimension as parameters, identification research for mechanical crack was carried out through the wavelet neural network method.