Research On Method To Exterior Qualitygrading Of Jujube Based On Machine Vision Technology

The jujube has the very high nutritional therapy health care value.In recent years,the interest in jujube industry is growing while the classification method is still in the primary stage of mechanical classification.In order to improve jujube product quality,it is essential to find an efficient and rapid method for the detection of jujube.In this thesis,Xinjiang jujube was used in the research,machine vision technology is used to classify the size,surface folds,surface diseases,cracks and other main exterior quality of jujube.The main research contents and conclusions are as follows:(1)Research on classification standard of exterior quality of Chinese Jujube.Concrete understanding present situation of Chinese jujube planting and postharvest processing through access to academic papers and survey. The grading standards of jujube was developed in postharvest processing, combined with the advantages of machine vision in jujube classification.(2)Research on the method of the size and fold of jujube classification based on machine vision. Normal jujube samples real transverse diameters were measured and then the image is carried on image segmentation, scanning the segmented image to detect jujube short axis of the pixel value; Using the least square method to fit the real value of the short axis and the pixel value, according to the classification standard of the true value and the function of the fitting, jujube size grading model was established, the correct rate of classification was95.65%. Using a gray level co-occurrence matrix to describe the surface fold texture of jujubes, analyzing the influence of structure factors for jujube fold surface texture parameters,three tectonic factor of jujube surface texture gray co-occurrence matrix were determined.Getting texture characteristic parameters of the generating direction from 0 degrees, 45 degrees, 90 degrees, 135 degrees and average value were took, generation step was 12 pixel values and image gray level was 64 level. Obtaining the contrast degree, correlation, energy,consistency, entropy, grayscale standard difference six parameters as jujube texture feature,the grading model of jujube fold was established based on BP neural network and SVM, the correct rates of classification were 95.23% and 94.29% respectively.(3)Research on the method of the disease and crack of jujube classification based onmachine vision. Firstly, on the basis of the difference in the tone value between the diseased and non-diseased areas in the H diagram, diseased area was extracted, and the extracted disease area to total surface area ratio was used as the threshold to achieve a high precision in identifying the diseased area. The test results of 163 diseased red jujubes and 500non-diseased jujubes showed that more than 16.87 mm2 diseased area could be correctly identified, accounting for 3.3% of the projected area of a jujube. The rates of correct recognition for the training set and the test set were 92.60% and 91.58%, respectively. To further improve the accuracy, the extracted diseased area to the surface area ratio was converted to the binary format. Combining with the mean and variance of color features of the jujubes, an SVM(support vector machine) model was developed to detect jujube diseases.The correct detection rate was 95.77 % for the training data set and 95.79 % for the test data set. In the I diagram, Otsu’s threshold method was first used to segment the regions on jujube surface, and then statistical and morphological methods were used to segment the crack regions and generate binary images. Using the invariant of cracks in the crack binary images,a crack recognition model was established. The adequacy of the model was tested on a data set of 500 samples, including 148 cracked jujubes and 352 non- cracked jujubes. For training data set, the detection rate was 94.9%. For the test data set, the detection rate was 94.55%.(4)Design of automatic grading system for jujube exterior quality based on machine vision.To present the overall design scheme of the system, familiar with MATLAB/GUI handle and callback function, according to the different function,four modules were divided.A auto grading system based on machine vision dates the appearance quality was designed whit friendly interface, Realizing the loading of jujube image, classification of jujube size,disease, folds and cracks, finally, the classification results were sorted.