【摘要】 为实现塑料垃圾的高质量回收利用,提出了一种基于Fisher向量(FV)与深度卷积神经网络(CNN)相结合的塑料垃圾精细分类方法。以塑料垃圾近红外光谱图为原数据,将深度CNN与FV编码相结合建立了塑料垃圾精细分类模型。针对几种典型塑料材料设计了验证实验,实验结果表明:该模型分类精度达到了91.25%,使塑料垃圾精细分类成为可能。更多还原

【Abstract】 In order to realize the high-quality recycling of plastic waste, a fine classification method of plastic waste based on Fisher vector(FV)and deep convolutional neural network is proposed.The method uses the near-infrared spectrum of plastic garbage as the original data, and combines the deep convolutional neural network with Fisher vector coding to establish a fine classification model of plastic garbage.The verification experiment is designed for several typical plastic materials.The experimental results show that the classification accuracy of the model reaches 91.25 %,which makes the fine classification of plastic waste possible.更多还原