Calculation Method Of Forest And Fruit Planting Area In Aksu Area Based On Remote Sensing Image Research And Implementation

Nowadays,remote sensing technology has broad applications in agriculture.Using satellite remote sensing technology for large-scale agricultural surveys has become an important application direction of remote sensing technology.The Aksu region in Xinjiang has a superior geographical environment for growing fruits.The characteristic forest and fruit industry has also become an important industry in the Aksu region.The traditional statistical methods for measuring crop area in the field of planting information statistics have limitations such as time-consuming,labor-intensive,and poor timeliness.The application of satellite remote sensing technology can effectively solve the problems of traditional methods,and can more easily grasp the information status of large-scale characteristic forest fruit resources,thereby accelerating the informationization of the forest fruit industry in Aksu.At the same time,the remote sensing identification of forest fruit plantation has very practical significance for monitoring and forecasting the development of Aksu's characteristic forest and fruit industry.In this experiment,the Aksu region was used as the research area,and five kinds of local forests,including red dates,apples,fragrant pears,walnuts and apricots,were taken as the research objects.First,collect sample database through field investigation and sampling,download related Landsat8 satellite remote sensing data information as experimental materials;secondly,perform pre-processing operations such as radiometric calibration,atmospheric correction,image mosaic and vector cropping on the remote sensing information;Three different supervised classification methods such as support vector machine,maximum likelihood method,and neural network method are used to classify the forest and fruit,and the classification results are subjected to three small patch removal treatments to perform precision analysis to select the optimal classification combination,And statistical information on the planting area of special forests and fruits;finally build a statistical system of planting area of special forests and fruits in Aksu.The main findings are as follows:(1)Experiments were conducted to supervise and classify the five characteristic forest fruits in Landsat 8 image respectively.Three types of methods including neural network method classification,maximum likelihood method classification and support vector machine method classification were selected.The overall accuracy and Kappa coefficient have reached 85%and more than 0.81.Achieved a better classification of five characteristic forest fruits in Aksu.(2)In order to improve the classification accuracy in the test area during the experiment,Majority Minority analysis,Clump and Sieve were used to perform three types of plaque removal processing on the three supervised classification results,and finally passed the accuracy In evaluation,the combination of Majority analysis after selecting support vector machine classification is the optimal classification processing method.The overall classification accuracy reaches 90.39%,and the Kappa coefficient is 0.8789,which effectively improves the classification accuracy.(3)Through the optimal classification combination in the experiment combined with the classification statistics module in the ENVI software,the planting area statistics of walnuts,red dates,apples,apricots and fragrant pears in the study area were carried out.The experimental statistical results are as follows.301006.383hm~2,of which walnut planting area is147627.6596hm~2;red jujube planting area is 147627.6596hm~2;apple planting area is22644.6809hm~2;apricot planting area is 15882.9787hm~2 and fragrant pear planting area is12403.1915hm~2.(4)Using.NET as the framework,ArcGIS Server as the GIS platform,and Internet Information Service(IIS)as the web server,we designed and built Aksu area economic forest fruit tree planting area statistics that support image display,distance and area measurement,and planting area statistics.The system has completed the research and implementation from the area measurement method.