基于图像分割和卷积神经网络的苹果分类机器人分类模型

Classification model of apple classification robot basedon image segmentation and convolutional neural network

为实现苹果分类机器人对苹果的高效分类,构建了基于图像分割和卷积神经网络的苹果分类新模型。该模型首先利用计算机视觉技术中的图像分割实现苹果图像中每个苹果的逐一分割,然后将原始图像和所有分割图作为完整图苹果分类模型和分割图苹果分类模型的输入,最后将完整图苹果分类模型和分割图苹果分类模型的输出集成,实现苹果的快速高效分类。仿真测试结果表明,所构建苹果分类模型对苹果数据库中苹果的分类准确率高于其他两种分类模型;该模型对测试数据集中苹果的分类准确率达到95.69%,对多种苹果的检测精度、召回率、F1分数达到100.00%。

A novel apple classification model based on image segmentation and convolutional neural network is constructed to realize the efficient classification of apples by apple classification robot.Firstly,the constructed model uses the image segmentation in computer vision technology to segment each apple in the original apple image.Then,the original image and all segmented apple images are taken as the inputs of the apple classification model based on the original apple image and the apple classification model using cropped apple images respectively.Finally,the outputs of the above two type apple classification models are integrated to realize fast and efficient classification of apples.The simulation results show that the apple classification accuracy of the constructed model is higher than that of the other two apple classification models.In addition,for the constructed apple classification model,the classification accuracy of apples in the test data set reaches 95.69%,and the classification precision,recall rate and F1 score of several apple varieties reach 100.00%.