双通道卷积神经网络模型电力设备图像识别

Research on Electric Equipment Image Recognition based on Double CNN and Random forest classification method

针对传统电力设备图像识别方法精度低、处理能力差的不足,提出一种基于双通道卷积神经网络(CNN)模型和随机森林分类的电力设备图像识别方法。在特征提取方面,双通道CNN模型通过两个独立的CNN模型来提取电力设备的特征;在识别算法方面,借鉴传统机器学习方法的优势,结合随机森林的优点,提出了结合深度学习的随机森林分类方法。最后,利用所提出的双通道CNN模型和随机森林分类方法对各类电力设备的图像进行了分类,实验表明所提出的方法可以有效地应用于各类电力设备的图像识别,并大大改善电力设备图像识别的识别率。

In allusion to the drawback of low accuracy rate and poor recognition ability of traditional electric equipment image recognition methods, an electric equipment image recognition method based on DCNN and random forest classification was proposed in this paper. In terms of feature extraction, DCNN was used to extract the features of electrical equipments by two independent CNN. In terms of recognition algorithm, a random forest classification method combined with deep learning was proposed considering the advantages of traditional machine learning methods and random forest. Finally, the proposed image recognition method based on DCNN and random forest was used to classify images of various electrical equipments. Tests have shown the effectiveness and higher recognition rate of the proposed image recognition method on electric equipment image recognition.