基于系统工程学的土壤侵蚀高分遥感监测与评价工程化模式

High-resolution remote sensing monitoring and evaluation engineering mode of soil erosion based on system engineering

高分遥感技术是开展土壤侵蚀监测与评价的重要手段,但目前应用的整体效能和水平不高。本研究基于系统工程学原理和方法,主要从工程化知识库构建、工程化遥感信息提取算法、工程化综合集成3个关键要素方面,提出了一种土壤侵蚀高分遥感监测与评价工程化模式,并选取陕西省横山县进行了示范应用实践。结果表明:高分遥感工程化自动总体分类精度87. 58%,经过工程化修正模式修正,分类精度达到95%,基于工程化应用系统平台与基础集成数据,整体监测与评价工作效率提升了2～3倍。该模式实现了土壤侵蚀高分遥感监测与评价过程环节的系统化和协同化,有效提升了监测与评价工作的稳定性和可靠性,为全覆盖、多频次推进土壤侵蚀监测与评价工作提供有效的技术模式参考。

[Background] High-resolution remote sensing technology is an important means to monitor and evaluate soil erosion. The regional applicability of high-resolution remote sensing information extraction automation algorithm is poor,the multi-period dynamic error fluctuation of remote sensing information extraction results is large,the integration of remote sensing information extraction and soil erosion monitoring and evaluation business is poor,and the overall application efficiency is not high.[Methods]Based on the research status of soil erosion monitoring and evaluation by high-resolution remote sensing,this paper considered the application of remote sensing technology in soil erosion monitoring and evaluation as a whole by using the principles and methods of system engineering,and put forward the construction idea of engineering model of high-resolution remote sensing monitoring and evaluation of soil erosion. This paper also clarified the conceptual connotation and main characteristics of the model,and put forward the overall framework of the model,as well as the construction content and implementation approach of the key elements of the model. [Results]After research,the model mainly includes 3 core elements: engineering comprehensive knowledge base model,engineering remote sensing information extraction algorithm model and engineering integrated model. Among them,the engineering comprehensive knowledge base model is mainly based on the spatial management framework system of parallel integration of soil and water conservation zoning and county administrative divisions,which standardizes and solidifies business and expert knowledge;the engineering remote sensing information extraction algorithm model is mainly to extract target objects and sub-regions for different high-resolution remote sensing information. An information extraction algorithm set with model and algorithm as the core is established,and the engineering integrated model mainly integrates the basic data,application software system and infrastructure equipment environment. In Hengshan county of Yulin city,Shanxi province,a demonstrating application on agriculture and animal husbandry was carried out. Automatic overall classification accuracy was 87. 58%. After the engineering model corrected,classification accuracy was95%. Based on the engineering application of the system platform and using integrated basic data,we completed the evaluation of soil erosion in Hengshan county,andthe overall efficiency of the evaluation increased by 2-3 times. [Conclusions]This model realizes systematic and collaborative optimization of monitoring and evaluating soil erosion using remote sensing,greatly improves its the stability,accuracy,practicability and comparability,and may provide technical support to promote dynamic assessment and governance of Chinese soil erosion.