Design And Implementation Of Apple Surface Defect Detection System Based On Deep Learning

At present,apple surface quality inspection is usually carried out manually,but this inspection method is time-consuming and inefficient.With the development of artificial intelligence,Deep learning algorithms has excellent ability of feature extraction and classification.By establishing a deep learning model,automatic classification of defects can be realized.To the above problems,this paper designs a deep learning-based system for detecting defects on apple surface.The research work of this paper can be summarized as follows:1)The experimental data is obtained and preprocessed.Firstly,the images of five types of apples,including health,mechanical injury,diseases and pests,wrinkle and decay,are collected by using web crawler and field sampling methods;Secondly,python algorithm is used to perform data augmentation operations on the sample images to expand the data scale and achieve data balance;Finally,these images are made into an experimental dataset for model training and testing.2)An apple surface defect detection algorithm based on lightweight convolutional neural network is proposed.Firstly,we construct AlexNet,VGG,ResNet,MobileNet network models for experiments and analysis.Secondly,we select AlexNet as the basic model and introduce the depthwise separable convolution to replace the standard convolution operation in the original network to extract image features.Then,we use the Leaky ReLU activation function instead of the ReLU activation function to perform a nonlinear mapping of the convolved results.Finally,we use the global average pooling method to replace the full connection layer in the original network,so that multiple feature maps output by the convolution layer are mapped in their units to obtain feature points.The experimental results show that the improved lightweight convolutional neural network not only reduces the number of model parameters and training time,but also improves the detection accuracy and speed of the model.3)The apple surface defect detection system is developed.The PyQt5 framework and python language are used to implement the specific functional modules according to the overall design scheme of the system.The system is mainly divided into user interaction module and defect detection module.The user interaction module is responsible for managing the identity information of users and administrators,and the defect detection module is responsible for detecting the input apple image and outputting the classification results.