Design And Research Of Wheat Imperfect Grain Detection System Based On OpenCV And GUI

Wheat was one of my country's reserve grains.In the process of wheat storage,market circulation,imperfect grains were important indicators to measure wheat quality.They were of great practical significance to do a good job in the detection of imperfect grains in wheat for important safe storage and market circulation of wheat.At present,the identification of imperfect grains of wheat mainly depended on the sensory inspection of inspectors.In the research of wheat imperfect grain detection based on image processing and machine learning,most of them were processed with neatly arranged wheat images.Although a small amount of data could quickly train a classification detection model with high recognition,it did not have wide practicability.This research applied Open CV image processing and deep learning to the detection of imperfect grains of wheat.The free-falling image acquisition method not only effectively increased the spatial randomness of wheat grains,but also conformed to the actual scene detection and recognition of imperfect grains of wheat.And the visual function of GUI(Graphics User Interface)in order to design and research the wheat imperfect grain detection system,which could provide a theoretical basis for the research and development of miniature and portable equipment for the intelligent detection of wheat imperfect grains.The main research contents and conclusions were as follows:(1)A double-layer analysis tray for sensory inspection of imperfect grains was designed to assist in the selection of imperfect grains of wheat.In the image acquisition,a simple device for collecting images of wheat was designed and developed,and the wheat falls freely onto the collection platform through the honeycomb blanking device,so that the collected wheat image grains were distributed randomly.(2)The Python-Open CV image processing technique was applied to wheat imperfection grain image processing,the operation results of different processing functions in Python-Open CV image enhancement,image morphology processing,and image segmentation operations were comprehensively compared,so as to select the optimal processing functions: mean filter function,Otsu's binarization function,and open operation function,and used Open CV image scaling function to process wheat seed images,enhanced image features,and established wheat imperfection grain image database.(3)Based on the image of imperfect wheat grains obtained by Python-Open CV image processing technology,the VGG16 deep learning model was established for the training and testing of imperfect grains of wheat.The recognition accuracy of imperfect grains followed the training data and iteration of the deep learning model.The number of times increased gradually.The neural network model obtained in this study had two types of recognition results for imperfect wheat grains.One was to classify and recognize perfect grains and imperfect grains.The overall recognition accuracy rate was as high as 85.4%;Species were identified according to the whole category of perfect grains and imperfect grains,and the overall recognition accuracy rate was as high as 92.7%.(4)Based on the research of image processing and deep model recognition,the graphical interface of the wheat imperfect grain detection system was designed and developed by using the Python-Tkinter function library to visualize the results of the wheat image processing and detection and recognition process.