| 摘 要: 针对外来物种入侵对湿地生态环境构成的严重危害和潜在威胁，开发了湿地外来入侵物种监测系统。本系统基于Linux操作系统，选用Eclipse开发软件、MySQL数据库和Openlayers地理信息系统，应用人工智能百度AI开发平台的图像识别接口服务，实现了外来入侵物种名录管理、物种识别、物种上报等功能。经过测试，本系统识别准确度较高，对终端性能要求较低，运行基本正常稳定，具有较强的移植性，适用于相关监测系统识别功能的二次开发，可以有效支持湿地外来入侵物种监测工作。 |
| 关键词: 图像识别；人工智能；监测系统 |
|
中图分类号: TP391
文献标识码: A
|
|
| Design and Implementation of a Wetland Invasive Species Monitor System based on Image Recognition |
|
YAN Ruihua
|
(Information Center of National Forestry and Grassland Administration, Beijing 100714, China )
yanruihua@buaa.edu.cn
|
| Abstract: This paper proposes to develop a wetland invasive species monitor system in view of the serious harm and potential threats to the wetland ecological environment caused by the invasion of alien species. The system adopts the Linux operating system, uses Eclipse development software, MySQL database and Openlayers geographic information system. It also utilizes image recognition interface services of Baidu AI (Artificial Intelligence) development platform. System functions, such as invasive species information management, species recognition and species reporting, are realized. Test results show that the system has high recognition accuracy, low terminal performance requirements, basically normal and stable operation, and strong transplantability. It can be applied to the secondary development of the recognition function of related monitor systems and can effectively support monitoring work of invasive alien species in wetlands. |
| Keywords: image recognition; artificial intelligence; monitor system |
