Research On The Extraction Technology Of Jujube Tree Planting Area Based On Time-series Satellite Remote Sensing Images

Alar Reclamation area is located in the edge of Taklimakan Desert,which is recognized as "fruit eugenic zone".In recent years,with the vigorous development of jujube planting industry in this area,it has gradually become one of the main planting areas of jujube in Xinjiang.The acquisition of jujube planting area in Alar Reclamation Area has important practical significance to provide relevant data support for corps crop decision-making.The use of remote sensing technology for jujube area extraction,greatly reduce the cost of measurement,but also ensure the objectivity and accuracy of the data.Taking Alar Reclamation Area as the research area and using Landsat 8 as the main data source,this paper studies how to extract jujube planting area quickly and efficiently from single temporal and time series images:1.Based on the single phase Landsat 8 images of jujube phenology,the method of jujube planting area extraction in Alar Reclamation area was studied.Before extracting the planting area of jujube trees in the study area,in order to better distinguish jujube trees from other crops,the spectral characteristics of jujube trees and other crops in the same period were analyzed,and the images with large difference in crop curve were selected,and then different supervised classification methods were used to extract the planting area of jujube trees.The effect of random forest classification method is better than other classification methods,and the classification result of jujube mature image is the best.2.Based on the Landsat 8 oli data of jujube growth cycle time series,taking alar city of Xinjiang as an example,the dynamic time warping algorithm was used for crop classification and recognition to explore the recognition ability of jujube using medium and high resolution time series images and DTW method.Through the experimental analysis,the DTW classification method based on NDVI has the highest accuracy,and the overall accuracy is 95.23%,The classification accuracy meets the needs of users.Therefore,it is feasible to extract the planting area of jujube in Alar Reclamation Area by using the classification method of temporal feature fusion proposed in this paper.3.Using the classification method of temporal feature fusion proposed in this study,the jujube planting area from 2013 to 2018 in Alar Reclamation area was extracted.Using years of data to verify the method,referring to the statistical yearbook over the years,the extracted jujube planting area can reach about 97%,which proves that the method can effectively extract the jujube planting area in the study area,and puts forward a new idea for the extraction of crop planting area.