基于移动终端的稻田飞虱调查方法  

A survey method based on mobile terminal for rice planthoppers in paddy fields

作　　者：俞佩仕 郭龙军 姚青[1] 杨保军 唐健 许渭根 陈渝阳 朱旭华 陈宏明[5] 张晨光 段德康 贝文勇 彭晴晖 YU Pei-Shi;GUO Long-Jun;YAO Qing;YANG Bao-Jun;TANG Jian;XU Wei-Gen;CHEN Yu-Yang;ZHU Xu-Hua;CHEN Hong-Ming;ZHANG Chen-Guang;DUAN De-Kang;BEI Wen-Yong;PENG Qing-Hui(School of Information Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China;State Key Laboratory of Rice Biology, China National Rice Research Institute, Hangzhou 310006, China;Plant Protection and Quarantine Bureau of Zhejiang Province, Hangzhou 310020, China;Zhejiang Tuopu Yunnong Science and Technology Co. , Ltd. , Hangzhou 310020, China;Xiangshan County Agricultural Technology Promotion Center Plant Protection Station of Zhejiang Province, Ningbo, Zhejiang 315700, China;Longyou County Agricultural Bureau Plant Protection and Quarantine Station of Zhejiang Province, Quzhou, Zhejiang 324400, China;Wan′an County Plant Protection and Quarantine Station of Jiangxi Province, Ji′an, Jiangxi 343800, China;Zhaoping County Disease and Biology Station of Agricultural Bureau of Guangxi Zhuang Autonomous Region, Hezhou, Guangxi 546800, China;Shaodong County Plant Protection and Quarantine Station of Agricultural Bureau of Hunan Province, Shaoyang, Hunan 422800, China)

机构地区：[1]浙江理工大学信息学院,杭州310018 [2]中国水稻研究所水稻生物学国家重点实验室,杭州310006 [3]浙江省植物保护检疫局,杭州310020 [4]浙江托普云农科技股份有限公司,杭州310015 [5]浙江省象山县农业技术推广中心植保站,浙江宁波315700 [6]浙江省龙游县农业局植物保护检疫站,浙江衢州324400 [7]江西省万安县植保植检站,江西吉安343800 [8]广西壮族自治区昭平县农业局病虫测报站,广西贺州546800 [9]湖南省邵东县农业局植保植检站,湖南邵阳422800

出　　处：《昆虫学报》2019年第5期615-623,共9页Acta Entomologica Sinica

基　　金：国家”863”计划项目(2013AA102402);浙江省公益性项目(LGN18C140007)

摘　　要：【目的】建立一种基于移动终端的稻田飞虱调查方法,以减轻测报人员劳动强度,提高稻田飞虱调查的客观性,实现稻飞虱调查结果可追溯。【方法】利用Android相机、可伸缩手持杆和装载控制相机APP的Android手机研制了稻田飞虱图像采集仪。在Android开发环境下,利用socket通信和视频编码等技术,实现Android相机的视频采集与编码模块、视频传输模块和相机命令控制模块等。利用Android NDK开发和Java web等技术,实现手机端的视频预览模块、手机控制模块、图像上传模块等。相机实时拍摄的视频将压缩成H.264格式,通过RTSP/RTP协议控制其传输至手机端。手机端通过解压缩,实现实时预览相机所拍摄的视频,并控制相机拍摄水稻茎基部飞虱图像,同时将图像传输到手机端。稻飞虱识别算法部署在云服务器上。手机端可选择稻飞虱图像上传至云服务器,云服务器运行稻飞虱自动识别算法,结果返回至手机端。【结果】基于移动终端的稻田飞虱调查方法利用手机可以实时预览相机拍摄的水稻茎基部飞虱画面,控制相机拍照。云服务器上稻飞虱自动识别算法对图像中的飞虱平均检测率为86.9%,虚警率为11.2%;对稻飞虱各虫态平均检测率为81.7%,虚警率为16.6%。【结论】基于移动终端的稻田飞虱调查方法可以便捷地采集到水稻茎基部飞虱图像,实现稻田飞虱不同虫态的识别与计数。该方法可大大减轻测报人员的劳动量,避免稻飞虱田间调查的主观性,实现稻飞虱田间调查的可追溯。【Aim】 This study aims to design and develop a survey method based on mobile terminal for rice planthoppers in paddy fields, so as to reduce the labor intensity, improve the objectivity of rice planthopper survey in paddy fields and trace the survey results of rice planthoppers.【Methods】 An image acquisition device for rice planthoppers in paddy fields was developed by an Android camera, a handheld extendable pole and an Android mobile phone loaded with camera APP. In the Android development environment, the video capture and encoding module, the video transport module and the camera controlling module of the Android camera were developed by the socket communication and video coding technologies. The video preview module, the mobile controlling module and the image uploading module on the mobile phone were developed by the Android NDK and Java web technologies. The video in the camera was compressed into H.264 format file in real time, which was transmitted to the mobile phone by the RTSP/RTP protocol. The phone can preview the video captured by the camera by decompression, control the camera to capture the rice planthopper images, and simultaneously transmit the images to the mobile phone. The automatic identification algorithm for rice planthoppers was deployed in the cloud server. The clients can select the rice planthopper images to upload to the cloud server. The identification results would be returned to the mobile phone after running the identification algorithm in the server.【Results】 By using the survey method based on the mobile terminal for rice planthoppers in paddy fields, the mobile phone can preview the video of rice planthoppers at the base of rice stem captured by the camera in real time and control the camera to take pictures. The automatic identification algorithm for rice planthoppers in the cloud server can identify both the rice planthoppers in the image and their developmental stages, with the average detection rates of 86.9% and 81.7%, and the false alarm rates of 11.2% and 16.

关 键 词：稻飞虱 田间调查 图像识别 自动计数 ANDROID手机 Android相机 

分 类 号：S435.112[农业科学—农业昆虫与害虫防治] S126[农业科学—植物保护]