基于KPCANet的工件图像特征提取研究

Research of Workpieces Image Feature Extraction Based on KPCANet

为了获得较高的分类准确率,从机器视觉在工业上的应用出发,利用一种简单的基于深度学习的图像特征提取框架KPCANet提取静止状态下的工件图像特征并进行分类。实验结果表明,与Bo W模型和PCANet框架相比,该算法能达到更高的工件分类准确率。

A simple deep learning network called KPCANet was applied to extract static workpiece image features for higher classification accuracy. Numerical experiments results showed that comparing with Bag of Words（ BoW） model and PCANet,KPCANet achieved a higher classification accuracy rate on workpiece image dataset.