基于人工神经网络和机器视觉的棉花分拣系统研究
Research on Cotton Sorting System Based on Artificial Neural Network and Machine Vision
朱西方1
作者信息
- 1. 河南工业职业技术学院,河南 南阳 473000
- 折叠
摘要
首先,介绍了卷积神经网络的原理,并基于双目视觉搭建了棉花分拣视觉系统;然后,基于 3×3 窗口、Sobel和Hough等算法,实现了棉花图像的边缘检测和特征提取功能;最后,基于卷积神经网络对棉花图像进行特征提取和优劣分类,并利用双目视觉对识别的棉花进行空间定位.实验结果表明:棉花分拣系统的准确率为 96.50%,能够有效地满足实际应用的要求.
Abstract
It firstly introduces the principle of convolution neural network,and then builds a cotton sorting vision system based on binocular vision.And then it realizes the function of edge detection and feature extraction of cotton image based on 3×3 Windows,Sobel,Hough and other algorithms.Finally,based on convolution neural network,the cotton image is feature extraction and classification of advantages and disadvantages,and the recognized cotton is spatially located by binocular vision.The experimental results show that the accuracy of this cotton sorting system is 96.50%,which can effectively meet the requirements of practical application.
关键词
棉花分拣系统/卷积神经网络/双目视觉/Sobel/HoughKey words
cotton sorting system/convolution neural network/binocular vision/Sobel/Hough引用本文复制引用
出版年
2025
国家科技期刊平台
NSTL
万方数据