Research On Grain Broken Methods Based-on Microscopic Images Processed

Rice is the survival of the human food crops that provide60%-70%of the energy needed for living people.As one of the major food crops, more than60%of our population is the staple food rice. Rice grains crack is produced rice, mildew, and the main reason for decline in food quality. Its cracks about the way their structural strength. The grain structure is the outward manifestation of the strength of the microstructure. Therefore, the discussion rice processing damage is significant from the microscopic point of view. Based on the macroscopic mechanical properties, microstructure and combined image processing technology and pattern recognition technology to research breakage way. The main contents are as follows:1. This text is in the United States PTC’s TMS-PRO Texture Analyzer to perform different damaged methods analog processing test for indica. Including shear, compression and bending mechanical properties test. Due to the special nature of the test sample, the bending fixture is designed by myself. The characteristic differences of these three different breakage force distribution mainly in this: the maximum damage force under compression distribution is unconcentrated span the range of large. The maximum damage force under shear distribution followed the compression. The maximum damage force under bending distribution is concentrated span the range of small. Compared non-drying and drying rice in the shear mode, to prepare for the subsequent recognition.2. Set-up platform for machine vision, Mechanical properties of the rice section of the above tests for image acquisition and image processing by MATLAB software. Microscopic image binarization, gray image segmentation, background removal, expansion and corrosion, skeleton extraction. Crack parameters extracted binary image. After image grizzled processing obtained texture parameters and crystallization point area ratio parameters, to prepare for the subsequent recognition.3. On the basis of image processing on the image were extracted texture parameters and shape characteristic parameters, such as:the average gray, average contrast, entropy, smoothness, contrast, energy, relevance, consistency, identity and third moment, spot area ratio, crack. Characteristic parameters with significant differences were obtained by scattergram and analysis of variance. Including the average gray level, the average contrast, entropy, smoothness, contrast, spot area ratio and crack. With significant differences of characteristic parameters obtained by scattergram and analysis of variance Significant. It’s used to identify ways of indica processing damage. Characteristic parameters with significant differences include: the average gray, average contrast, entropy, smoothness, contrast, spot area ratio and crack.4. Significant characteristic parameters were identified and classified by artificial neural network. Which include the average gray, average contrast, entropy, smoothness, contrast, spot area ratio and crack. The topics were designed three kinds of identifiers. Which under different moisture crack average recognition rate of92.5%. In different force ways, the average recognition rate of crystallization point area ratio of90.2%. The average recognition rate of significantly texture parameters is87.5%.