Study On The Terrace Information Extraction Of Gan Qingning Loess Plateau Based On High-score Image

Loess Plateau region is one of the most fragile areas of ecological environment in China and its serious soil erosion has restricted the sustainable development of the region.Since large areas of terraced fields can change surface topography and water erosion,for a long time,terraced fields have played an important role in soil and water conservation as an important engineering measure in the Loess Plateau area of China.In the work of soil and water conservation,it is an important part of monitoring and evaluation of soil and water conservation to grasp the information of regional terraced fields in time.At present,the research methods of terraced fields are mainly based on statistics and small-scale remote sensing investigation.In this paper,a rapid and effective method for terrace information extraction from remote sensing images at large scale is explored,which provides a scientific basis for soil and water conservation and ecological construction in the Gan Qingning loess plateau region.In this paper,the Gan Qingning loess plateau region is taken as the research area and high resolution remote sensing images are chosen as the main data source to accomplish the extraction of terrace information at large scale with two object-oriented classification methods.The extraction results are compared and analyzed through subjective and objective ways and the extraction accuracy is evaluated.Finally,the applicability of the methods and the rationality of terrace distribution are discussed.The main work and conclusions of the paper are listed as follows:(1)After the analysis of terraces' characteristics and phenological data in the study area,high resolution remote sensing images are used to accomplish the remote sensing measurement of terraced field information.A rapid survey method using GF-1 WFV data for terrace measurement at large scale is proposed.(2)A set of extraction and classification rules is constructed during the process of extracting information with the rule-based method.It can fully mine all kinds of information and improve the accuracy of extraction.Sample-based method is simple and the extraction accuracy is high.However,both of these two methods have some shortcomings.(4)On the basis of the results of terrace extraction,the rationality of the distribution of terraced fields is analyzed by analyzing the precipitation,elevation,mountain slope and human factors in the study area.The main influencing factors of terrace distribution in Gan Qingning loess Plateau are precipitation,elevation,mountain slope and human factors.(3)Combined with survey data and related information,the extraction results were tested by subjective and objective evaluation.The results show that the average precision of rule-based method is as high as 73.03%,and the Kaapa coefficient is 0.6130;the average precision of sample-based method achieves 83.15% and the Kaapa coefficient is 0.6822.The average accuracy of the latter is higher than the former.(4)On the basis of the extraction results,the rationality of terrace distribution is analyzed.The main influencing factors of terrace distribution in Gan Qingning loess Plateau are precipitation,elevation,mountain slope and human factors.(5)Through the extraction of Gan Qingning terraced fields based on remote sensing,the distribution and scale of terrace in this area can be obtained.It is good for comprehensive understanding of the distribution and the amount of terraced fields in the area.It can provide reference for water and soil conservation and development of terrace furtherly.