基于知识推理的农作物空间分布监测方法研究

利用遥感技术进行农作物识别和监测是遥感应用领域的重要研究内容之一。以2006—2007年两个时相的CBERS-02 CCD影像为主要遥感数据源,对山东省某市的主要农作物的种植分布情况进行监测。将农作物的物候特征、光谱特征和纹理特征及GIS辅助信息等多源信息融合,建立识别知识规则,通过知识推理,逐步识别出冬小麦、夏玉米和棉花。最后,利用混淆矩阵对实验结果进行验证。通过分析比较,证明上述方法在监测作物空间分布方面具有较高精度。

The Research of Crop Spatial Distribution Monitoring Through Knowledge - Based Reasoning

Monitoring crops based on remote sensing technology is an important research direction in remote sensing applications.In this paper,two phases of CBERS-02 CCD images between 2006 and 2007 are selected to monitor the main crop spatial distribution in a city of Shandong Province.The characteristics of crop phenology,spectral,texture and auxiliary information of GIS are merged,and knowledge rules are established based on the above information by knowledge-based reasoning method-winter wheat,summer maize and cotton are identified step by step.At last,confusion matrix is taken to verify the monitoring result.The evaluation result shows that the method stated in this paper has high accuracy in crop spatial distribution.