| 采用AIS计算中西太平洋延绳钓渔船捕捞努力量 |
| |
| 引用本文： | 杨胜龙,张胜茂,周为峰,崔雪森,张忭忭,樊伟.采用AIS计算中西太平洋延绳钓渔船捕捞努力量[J].农业工程学报,2020,36(3):198-203. |
| |
| 作者姓名： | 杨胜龙 张胜茂 周为峰 崔雪森 张忭忭 樊伟 |
| |
| 作者单位： | 中国水产科学研究院东海水产研究所,农业农村部远洋与极地渔业创新重点实验室,上海 200090 |
| |
| 基金项目： | 国家重点研发计划（2019YFD0901405），中央级公益性科研院所基本科研业务费（2019T09）；国家自然科学基金项目（41606138）；农业部外海渔业开发重点实验室开放基金资助（LOF2018-01） |
| |
| 摘 要： | 对渔船捕捞行为和捕捞强度空间高分辨率的估计可以作为海洋资源管理和生态脆弱性评估的重要信息。为识别远洋延绳钓渔船作业状态,该文基于2017年10-11月中西太平洋延绳钓渔船卫星船舶自动识别系统(automatic identification system,AIS)数据和捕捞日志数据,采用支持向量机(support vector machine,SVM)学习方法,构建了中国中西太平洋延绳钓渔船捕捞作业状态(捕捞/非捕捞)分类模型。通过计算模型分类准确率、精确率、敏感度和特异度来评价模型对渔船作业状态分类能力。结果表明,模型训练数据的准确率为95.24%(Kappa系数为0.9),验证数据的准确率为93.85%(Kappa系数为0.87)。采用构建好的模型识别2017年10月和11月中西太平洋延绳钓渔船共计125624条AIS记录数据,模型准确率在83.3%(Kappa系数为0.67)。2017年10、11月所有数据分类精确率为82.33%,灵敏度为88.32%,特异度为77.27%。渔船主要作业空间在168°E^173°E,12°S^18°S,有3个明显的作业强度较高区域。基于SVM模型和日志记录的捕捞强度信息在空间上相关性很高(r>0.98),SVM模型识别的渔船捕捞努力量空间分布特征和实际吻合。捕捞努力量与单位捕捞努力量渔获量(catch per unit of effort,CPUE)、渔获尾数、渔获质量和投钩数的相关系数分别是0.68、0.93、0.93和0.94。基于AIS信息挖掘的渔船空间捕捞努力量可用于渔业资源分析。
|
| 关 键 词： | 支持向量机 模型 自动识别系统(AIS) 延绳钓 捕捞努力量 |
| 收稿时间： | 2019/8/13 0:00:00 |
| 修稿时间： | 2020/1/16 0:00:00 |
Calculating the fishing effort of longline fishing vessel in the western and central pacific ocean using AIS |
| |
| Yang Shenglong,Zhang Shengmao,Zhou Weifeng,Cui Xuesen,Zhang Bianbian and Fan Wei.Calculating the fishing effort of longline fishing vessel in the western and central pacific ocean using AIS[J].Transactions of the Chinese Society of Agricultural Engineering,2020,36(3):198-203. |
| |
| Authors: | Yang Shenglong Zhang Shengmao Zhou Weifeng Cui Xuesen Zhang Bianbian and Fan Wei |
| |
| Institution: | 1. Key Laboratory of Oceanic and Polar Fisheries, Ministry of Agriculture and Rural Affairs, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, 200090, China; 2. Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, Chinese Academy of Fishery Sciences, Shanghai 200090, China,1. Key Laboratory of Oceanic and Polar Fisheries, Ministry of Agriculture and Rural Affairs, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, 200090, China; 2. Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, Chinese Academy of Fishery Sciences, Shanghai 200090, China,1. Key Laboratory of Oceanic and Polar Fisheries, Ministry of Agriculture and Rural Affairs, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, 200090, China; 2. Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, Chinese Academy of Fishery Sciences, Shanghai 200090, China,1. Key Laboratory of Oceanic and Polar Fisheries, Ministry of Agriculture and Rural Affairs, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, 200090, China; 2. Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, Chinese Academy of Fishery Sciences, Shanghai 200090, China,1. Key Laboratory of Oceanic and Polar Fisheries, Ministry of Agriculture and Rural Affairs, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, 200090, China; 2. Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, Chinese Academy of Fishery Sciences, Shanghai 200090, China and 1. Key Laboratory of Oceanic and Polar Fisheries, Ministry of Agriculture and Rural Affairs, East China Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Shanghai, 200090, China; 2. Key and Open Laboratory of Remote Sensing Information Technology in Fishing Resource, Chinese Academy of Fishery Sciences, Shanghai 200090, China |
| |
| Abstract: | In order to provide relevant information for natural resource management, impact assessment and marine spatial planning, high-resolution estimation of longline fishing gear is needed. Because the use of satellite vessel monitoring system (VMS) data is limited by data access and receiving offshore ship location information. In this paper, based on the data of automatic identification system (AIS) and logbook data of Chinese longline fishing from October to November 2017, a model of fishing detection based on support vector machine (SVM) learning method is established to classify the fishing and non fishing activities in the central and Western Pacific Ocean, to carry out longline fishing and draw fishing effort map in the Western and Central Pacific Ocean. Before constructing a fishing detection model, each AIS point was classified and pre-labeled as potential fishing and non-fishing events by an expert based on information on information on fisheries characteristics as obtained from literature, analysis of the tracks and logbook data. The performance of the fishing detection model was evaluated by the overall accuracy, precision, sensibility and specificity. Fishing intensities were computed from the known fishing positions and the estimated fishing positions were compared with correlation coefficient. The spatial correlation coefficient of fishing intensities and catch data was computed to quantify the extent to which the distribution of longline vessels describes tuna distribution. The results showed that the overall accuracy and the Kappa coefficient of the model training dataset were 95.24% and 0.9, respectively. The precision, sensibility and specificity were 94.74%, 93.44%, 96.56%, respectively. The overall accuracy and the Kappa coefficient of the model testing dataset were 93.85% and 0.87 respectively. The precision, sensibility and specificity were 93.49%, 90.67%, 95.91%, respectively. Then the constructed model was used to identify all the AIS data for 12 longline fishing vessels in the Western and Central Pacific in October and November 2017, with an overall accuracy rate of 83.3% and the Kappa coefficient was 0.67. The precision, sensibility and specificity were 82.33%, 88.32%, and 77.27%, respectively. The longline fishing distinct in October and November 2017 mainly located in 168°E-173°E, 12°S-18°S, and there were three obvious high fishing intensity areas on the map. The spatial distribution of fishing effort was significantly different between October and November. The distribution of fishing intensity in November was closer to the island than in October and the value of fishing intensity in October was lower than November. The fishing intensity information based on the SVM model and logbook data records was highly correlated (r>0.98). The spatial distribution characteristics of the fishing effort of the fishing vessel identified by the SVM model were similar to the known fishing positions. But the fishing intensities were calculated from known fishing positions was higher than that of estimated fishing positions. The spatial correlation coefficients of cumulative fishing effort and catch per unit of effort (CPUE), catch tail, catch weight and number of hooks were 0.68, 0.93, 0.93 and 0.94, respectively. The fishing capacity of fishing vessels based on AIS information mining can also be used as an alternative method for fishery resource analysis. |
| |
| Keywords: | support vector machine models automatic identification system longline fishing effort |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《农业工程学报》浏览原始摘要信息 |
| 点击此处可从《农业工程学报》下载免费的PDF全文 |
|
