| 基于激光雷达的海滩垃圾快速识别 |
| |
| 引用本文： | 何钰滢,葛振鹏,李道季,施华宏,韩震,戴志军.基于激光雷达的海滩垃圾快速识别[J].海洋学报,2019,41(11):156-162. |
| |
| 作者姓名： | 何钰滢 葛振鹏 李道季 施华宏 韩震 戴志军 |
| |
| 作者单位： | 华东师范大学 河口海岸学国家重点实验室,上海,200241;上海海洋大学 海洋科学学院,上海,201306 |
| |
| 基金项目： | 国家重点研发项目（2016YFC1402202）；国家自然科学基金（41576087）。 |
| |
| 摘 要： | 全球日益增多的海滩垃圾，不仅造成海洋环境污染，严重威胁海洋生态系统健康，也对生物栖息地有着不可估计的影响。如何高效准确地对海滩垃圾进行监测和识别，是处置海滩垃圾过程的技术难点之一。基于此，本文以长江口南汇边滩为实验区，通过在海滩上设置常见垃圾样品，随后利用激光雷达记录的全波形数据和BP神经网络模型，以快速鉴别海滩垃圾类型。结果表明:基于激光雷达提取的垃圾全波形数据中回波振幅和回波宽度的差异，可用来识别海滩垃圾。构建的BP神经网络可有效将海滩垃圾分为泡沫类、布类、金属类、纸类及塑料类，最高识别率达到79%。此外，由于不同材质海滩垃圾的原材料成分存在相似或同质，会对精确识别区分垃圾类型造成一定的干扰，从而影响神经网络的识别率。可见，将激光雷达应用于识别海滩垃圾，为海滩垃圾的监测提供了新的方法。
|
| 关 键 词： | 激光雷达 BP神经网络 海滩垃圾识别 |
| 收稿时间： | 2018/9/22 0:00:00 |
| 修稿时间： | 2018/12/14 0:00:00 |
LiDAR-based quickly recognition of beach debris |
| |
| He Yuying,Ge Zhenpeng,Li Daoji,Shi Huahong,Han Zhen and Dai Zhijun.LiDAR-based quickly recognition of beach debris[J].Acta Oceanologica Sinica (in Chinese),2019,41(11):156-162. |
| |
| Authors: | He Yuying Ge Zhenpeng Li Daoji Shi Huahong Han Zhen and Dai Zhijun |
| |
| Institution: | 1.State Key Laboratory of Estuarine and Coastal Research, East China Normal University, Shanghai 200241, China2.College of Marine Sciences, Shanghai Ocean University, Shanghai 201306, China |
| |
| Abstract: | There is an increasing amount of beach debris worldwide which have serious impacts on the marine environment, especially to the marine ecosystem health and biological habitats. It has been one of the great technological difficulties on how to monitor and identify beach debris efficiently during the process of the accurately disposing beach debris. Therefore, in this paper, a new recognition method of beach debris was proposed based on field beach debris experiment on Nanhui Beach by combination of LiDAR (light detection and ranging) with record full waveform data and the Back Propagation (BP) neural network model. The results reveal that the echo amplitude and width extracted from full-waveform data can be used to identify beach debris because of their distinct waveform features. Meanwhile, beach debris can be effectively classified into foam, cloth, metal, paper and plastic with the highly accuracy rate of 79% by the BP neural network recognition. Moreover, it can be found that some beach debris are difficult to identify owing to the same material composition for these debris, which may disturb the recognition rate of BP neural network to great degree. Therefore, it can be expected that a new monitoring tool for beach debris identification by LiDAR will be popular in future. |
| |
| Keywords: | LiDAR BP neural network beach debris recognition |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《海洋学报》浏览原始摘要信息 |
| 点击此处可从《海洋学报》下载免费的PDF全文 |
|
