Recognition Method Of Solanum Melongena Leaf Disease Based On Image Process Research

China is the world's largest producer of Solanum Melongena. In recent years, with the expansion of planting area, Solanum Melongena leaf disease gradually become the main restricting factor of Solanum Melongena high yield, high quality and high benefit production, inseriously continuous croppingfield, even has reached the point that it cannot continue planting kinds of Solanum Melongena. Therefore, timely detect and accurately identify the Solanum Melongena diseases in early onset become critical. Now, the method of vegetable diseases recognition has certain limitation, meets the requirement of modern agriculture. Recognition method of vegetable diseases currently has some limitations, it is difficult to meet the requirements of modern agriculture. With the continuous development of computer vision and pattern recognition technology, the method of using computer to intelligently identify crop diseases arises at the historic moment. This paper fully understand the current research in this field at home and abroad, on this basis, by means of using computer vision technology, the combination of image processing and pattern recognition technology, emphatically analyzed the color, shape, texture characteristic feature parameters of the disease spot in Solanum Melonge, put forward Solanum Melongena leaf disease recognition method based on image processing. On the basis of the current domestic and foreign research in this field, by means of using computer vision technology, combined with image processing and pattern recognition technology, this paper focused on the color, shape, characteristic parameters of the disease spot in Solanum Melonge, put forward a recognition method of Solanum Melongena leaf disease based on image processing.In this paper, the main work and innovation points include the following aspects:(1) Integratethe Solanum Melongena disease image acquisition system which was suit to monitor the Solanum Melongena disease. Through analyzing and studying the image quality requirements of Solanum Melongena disease identification, the performance of acquisition equipment such as power and wireless communication distance, developed a set of Solanum Melongena disease image acquisition system with solar power and remote wireless communication independently.(2) Research the method of disease spot segmentation. After getting the disease image, do the pre-processing, such as image segmentation, image equalization, etc. According to the characteristics of crop disease images and sampling conditions, analyze and compare the effect of several kinds of traditional image pre-processing methods, modify algorithm, improve the image processing effect, preparing for the post-processing of the image. According to the characteristic that H-component diagram is concentrated in an area when put the leaf image in to HSI color space, namely disease spot Hue concentration, get the background section by H component image binarization processing, thus acquire disease spot image.(3) Research the method of feature extraction. Chose 12 statistical characteristics which consist of the first moment and second moment of each component of disease spot image in RGB and HSI color space as color features, calculate 11 statistical characteristics included disease spot circularity, rectangularity, eccentricity, shape complexity, and 7 Hu invariant moment as shape characteristics, the gray level co-occurrence matrix was used to calculate 8 statistical characteristic which comprise correlation, energy, contrast, entropy the average and variance of disease spot image as the texture features. A total of 31 characteristic parameters. Through variance and principal component analysis(pca) to select 20 classification ability of the characteristic parameters of feature vector.(4) Compared the result of three different pattern recognition classification method classifying test set. The test results show that the recognition accuracy of phomopsis vexans in Solanum Melongena reached 90%, which show that this method can rapid and accurate identification of Solanum Melongena leaf diseases, and to achieve real-time detection of disease of Solanum Melongena field in an open environment and provides technical support.