脐橙色泽与着色率的计算机视觉检测技术研究

Such defects as great intensity of labor, low productive efficiency, hard to realize the criteria of gradation and unsteadiness of the accuracy of gradation lie in the problem of manual gradation of the color and luster character of navel oranges. To conquer all the above defects, the article introduces how to process the graphs of navel oranges by adopting the technique of computer visions and makes a quantitative description of the qualitative color and luster of navel oranges on the basis of the analysis of the color characters of the fruit surface, and finally establishes the models of gradation through the neural network. The results of the experiment suggest that the consistency rate of the gradation model of the gradation of the color and lustre and manual gradation has reached above 90% or 92% respectively。