水果分拣机械手控制系统研究——基于BP神经网络和图像识别

Research on the Control System of Fruit Sorting Manipulator—Based on BP Neural Network and Image Recognition

首先,构建了水果分拣机械手系统整理框架,并对工业相机的选型和相机标定进行分析研究;然后,基于BP神经网络实现了对目标水果的识别;最后,利用图像处理的模板匹配对目标进行实时定位,实现了水果分拣。试验结果表明:系统可以实现对多种水果的分类,准确性非常高,具有可行性。

It firstly constructs the frame of sorting robot system,analyzes the selection and calibration of industrial cameras,then realizes the recognition of target fruit based on BP neural network.Finally,it realized the real-time positioning of the target by using the template matching of image processing to realize the fruit sorting manipulator system.The results show that the system can classify many kinds of fruits,and the accuracy is very high,which proves the feasibility and accuracy of the system.