A review of hyperspectral imaging-based plastic waste detection state-of-the-arts
International Journal of Electrical and Computer Engineering (IJECE)（2023）
摘要
Plastic waste issues emerged from the build-up of plastics that negatively impacts the environment. As a result, plastic waste detection is proposed in many research studies to tackle the problems. Therefore, this paper aims to review hyperspectral imaging techniques and machine learning in plastic waste detection. Hyperspectral imaging techniques are found to be effective in detecting plastic waste and microplastics as they were able to capture plastic reflectance spectral by using the near-infrared sensor. However, the review also shows that hyperspectral imaging techniques were less efficient in capturing the electromagnetic spectrum of black plastics due to carbon-black absorption properties. Carbon-black strongly absorbs light in the ultraviolet and infrared spectral range of the electromagnetic spectrum, therefore not detected by the near-infrared sensor. This paper also reviews how machine learning can alternatively detect and sort all types of waste, including plastics. Multiple studies show that the machine learning model achieved good accuracy in detecting all types of plastics based on the waste dataset. Finally, it can be seen that the spectral information of plastic can be used as feature extraction for machine learning models for better plastic detection. It is hoped that this study will contribute to more systematic research on the same topic.
更多查看译文
关键词
waste,plastic,imaging-based,state-of-the-arts
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要