| Panax notoginseng is a unique and precious Chinese herbal medicine and one of the priority agricultural products with high plateau characteristics in Yunnan Province.It is of great significance to grasp the planting situation of Panax notoginseng for guiding the development of local agriculture.With the development of machine learning,image processing,software engineering,remote sensing and other technologies,it is possible to monitor Panax notoginseng planting area in large area with long time series.At present,there is a lack of research on remote sensing monitoring of Panax notoginseng planting area.Landsat image data has become the main data source for crop remote sensing monitoring due to its advantages of long time sequence and high resolution.The main contents and conclusions are as follows:(1)Research on the extraction algorithm of the Panax notoginseng planting area considering hillshade.First,this study screened out Landsat images with less clouds in 2019,2015,2010,2005,2000,1995 and 1990.To eliminate the impact of the hillshade,the hillshade restoration effect of regression correction method,C correction,SCS+C correction,VECA correction and shadow compensation method were tested.Second,spectral features,texture features,topographic features and tasseled cap transformation features were constructed,and above features were optimized through the J-M distance method.Then,the experiment compared the extraction results of three classification methods:random forest,CART decision tree,and support vector machine,and selected the best results to extract Panax notoginseng planting areas in different years.The results showed that:The regression correction method can better restore the brightness value of the shadow area of the mountain in the remote sensing image.The features combination through J-M distance selection had higher sample distinguishability.The results of random forest classification had high classification accuracy,the overall accuracy of classification results in each year was above 0.91,and the Kappa coefficient was above 0.86.(2)Research on the algorithms of spatiotemporal change monitoring and driving force analysis of Panax notoginseng planting area.This paper analyzed the spatiotemporal changes of the Panax notoginseng planting area in Wenshan Prefecture,analyzed its changing trends in altitude,slope and aspect,and used the method of land use transfer matrix and dynamic degree to monitor the changes in the distribution of Panax notoginseng planting area.Then,correlation analysis and principal component analysis were used to analyze the driving force of the changes in the Panax notoginseng planting area.The study found that:The area of Panax notoginseng planting areas in Wenshan Prefecture in 2019,2015,2010,2005,2000,1995,and 1990 were 279.185 km~2,283.266 km~2,40.32 km~2,12.036 km~2,27.171 km~2,17.854 km~2,and 14.533 km~2 respectively.Panax notoginseng planting areas were mainly distributed in Qiubei County,Yanshan County and Wenshan City in Wenshan Prefecture.From 1990 to 2019,the Panax notoginseng planting area was mainly planted in areas with an altitude of 700-1900 meters and a slope less than 25,and there was a tendency to expand to the middle and low altitudes,the northern slope and the western slope.From 1990 to 2019,the area of Panax notoginseng planting area showed a trend of slow growth firstly,then slowly decreasing and then rapidly increasing,especially from 2010 to 2015,the area of Panax notoginseng planting area increased by 240.047 km~2.The variation of Panax notoginseng planting area in Wenshan Prefecture is affected by the factors of population,living standard of residents,climate,economic development level and industrial structure.(3)Data visualization development of Panax notoginseng planting area in Wenshan Prefecture.Based on JS language and GEE platform,this paper completed the application development of the distribution display of the Panax notoginseng planting area,and realized the distribution and statistical area display of the Panax notoginseng planting area in Wenshan Prefecture in each year. |