Chain-chip automatic sorting array method based on computer vision
Chain-chip automatic sorting array method based on computer vision
复制标题
基于计算机视觉的链片自动排序阵列方法
DOI：
10.1504/ijwmc.2017.10006557
复制
发表时间：
2017-07
影响因子：
--
通讯作者：
张利
中科院分区：
文献类型：
--
作者：
单晓杭;叶必卿;张利
链片的自动识别与分选是自动化生产线的重要技术之一。为了提高链片排列和整个生产过程的效率和自动化程度，本文提出了一种基于计算机视觉和人工神经网络的链片自动排序排列方法。首先，提出了链式芯片自动分选阵列机。然后，在Hu不变矩算法的基础上，对采集图像的不变矩进行分析，得到用于分类的特征参数。设计了神经网络分类器，实现了链式木片的自动分选技术。实验结果表明，前三阶不变矩可以作为链片识别的特征参数。该方法具有精度高、采样性能好、抗噪声能力强等优点。能够满足复杂环境下目标识别的要求。
The automatic identification and sorting of chain-chip is one of the important technologies in automatic production line. In order to improve the efficiency and automation of the chain-chips arraying and the whole production process, this paper provides a chain-chip automatic sorting array method based on computer vision and Artificial Neural Network (ANN). Firstly, the chain-chip automatic sorting array machine is proposed. Then, on the basis of Hu invariant moment algorithm, the invariant moments of the acquisition image are analysed to gain the characteristic parameters for sorting. The neural network classifier is designed to realise automatic sorting technology of chain-chips. Experiment results show that the first three-order invariant moment can be used as characteristic parameters for chain-chip identification. This method has the advantages of high accuracy, good sampling performance and strong anti-noise ability. It can meet the demands of object identification requirements in complicated environment.