摘要
研究植被生态自动化观测方法,进而确定仪器的组成结构,根据传感器的接口、协议类型和神经网络处理器的应用方式设计了基于ARM架构控制处理器的外设接口,进一步结合非RKNN模型的转换应用流程,设计了控制处理器数据采集、处理、存储、传输和控制交互的具体功能和流程。针对3种不同的植被类型,将仪器分别部署在内蒙古、广西和安徽3个国家级农业气象试验基地,采用仪器测量数据与人工平行观测的方法进行外场试验。结果表明,这种植被生态自动化观测方法是可行的、有效的,仪器性能指标也符合实际业务观测要求。
This paper studied the observation method of vegetation ecological automation,and then determined the composition and structure of the instrument.According to the sensor interface,protocol type and application mode of neural network processor,the peripheral interface of control processor based on ARM architecture was designed.Furthermore,combined with the application process of non RKNN model conversion,the specific function and process of the data acquisition,processing,storage and transmission,control interaction of control processor were designed.According to three different vegetation types,the instruments were deployed in three national agrometeorological experimental bases in Inner Mongolia,Guangxi and Anhui,and the field test was carried out by using the method of instrument measurement data and manual parallel observation.The results showed that the automatic observation method of vegetation ecology was feasible and effective,and the performance index of the instrument was also in line with the actual business observation requirements.
作者
王贝贝
姚艳丽
张争
王苗苗
郭琰煊
WANG Bei-bei;YAO Yan-li;ZHANG Zheng(Henan Zhongyuan Photoelectric Measurement and Control Technology Co.,Ltd.,Zhengzhou,Henan 450047;Xinjiang Xingnong Network Information Center,Urumqi,Xinjiang 830002;Hebei Meteorological Bureau,Shijiazhuang,Hebei 050000)
出处
《安徽农业科学》
CAS
2021年第17期201-205,共5页
Journal of Anhui Agricultural Sciences
关键词
植被生态自动化观测
深度学习
植被关键特征参数测量
Automatic observation of vegetation ecology
Deep learning
Measurement of key characteristic parameters of vegetation