多源数据小麦全蚀病区域尺度预警研究

项目名称： 多源数据小麦全蚀病区域尺度预警研究

项目编号： No.41501481

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 天文学、地球科学

项目作者： 郭伟

作者单位： 河南农业大学

项目金额： 20万元

中文摘要： 全蚀病是一种造成小麦严重产量损失的毁灭性病害，近年来在各麦区呈多发趋势。区域尺度的空间预警对病害防治有重要意义。目前小麦病害预测方法往往不考虑其它致病因素，仅靠气象数据在大范围进行，导致预测精度较低，无法指导实际田间操作。如何整合多源数据来获取导致病害发生的关键信息，如何实现面状连续的空间预警是病害预测要解决的关键问题。本项目从小麦全蚀病的发病机制出发，利用遥感数据（面状连续）、传感和气象数据（时相连续）及其他环境数据提取和解析导致病害发生的三个关键要素：病原区、现势小麦长势和农田环境信息，在此基础上结合农学和植物保护知识获取病害发生的空间及时序特征，最终构建病害因子信息库，以此为驱动信息，耦合元胞自动机与多智能体系统模型和图像分割方法，构建面向对象的小麦全蚀病空间预警模型。研究结果能从理论和方法上提高区域尺度小麦全蚀病的预测水平。

中文关键词： 病虫害；多光谱遥感；高光谱遥感；作物长势监测；植被遥感建模

英文摘要： Wheat take-all is one of the devastating diseases threatening wheat production. Space early warning in a regional scale is of significant importance to disease control. As existing disease forecasting conducted within a wide range of areas uses only meteorological data without taking into consideration other pathogenic factors, the forecasting results are not accurate enough and of little referential significance to practical field operation. Hence, the key problems to be solved by disease forecasting are how to acquire key information on wheat take-all disease by using multi-source data, and how to get space-continuous warning in a regional scale. Based on the pathogenesis of wheat take-all and with the application of remote sensing data (area-continuous), sensing and meteorological data (temporal-continuous) and other environmental data, this research is designed to extract and analyze three key factors of wheat disease: the region being infected by take-all disease last year, existing growing information of wheat and environmental information of farmland. On this basis, the spatial and time-sequence characteristics of wheat take-all disease were observed by combining with the knowledgement of agronomy and plant protection. We finally built the database of wheat take-all pathogenic factor. Disease pathogenesis information database as drive information, the method of CA - MAS model and image segmentation was adopted to build the object-oriented wheat take-all disease early warning model. The results could improve the forecasting of wheat take-all disease in a regional scale theoretically and methodologically.

英文关键词： Crop Disease;Multi-spectral Remote Sensing;Hyperspectral Remote Sensing;Crop Growth Mornitoring;Modeling of Vegetation by Remote Sensing