Analysis Of The Evolution Of GF-1 And Sentinel Images Of The Crop Planting Structure Based On Time And Space

As the hometown of high-quality soybeans and high-starch corn in China,hailun city is an important commodity grain base and main crop producing area in the country.To understand the space-time change characteristics of the main crop planting areas in this area,and to analyze the driving force change,to master the situation of agricultural production using agricultural production resources within the space,and to understand important information such as crop types,planting structure,distribution characteristics,etc.,so as to provide scientific basis for crop structure adjustment and optimization.Crop planting structure monitoring and yield estimation are key areas of precision agriculture remote sensing,and the research results are of great significance for guiding crop planting structure adjustment and formulating agricultural policies.This paper takes Helen City in Heilongjiang Province as the research area,uses Sentinel and multi-temporal GF-1 as remote sensing data sources,uses sample-based farmland information extraction methods,and reasonably sets the scale level and merge level.The image is segmented according to the brightness,texture and color of adjacent pixels,and the edge segmentation algorithm is adopted.This algorithm is fast to calculate and can generate multi-scale segmentation results with only one input parameter.Secondly,through the difference control of the boundary on different scales,a multi-scale segmentation from fine to coarse is generated,the farmland information in the study area is extracted,and then an object-oriented decision tree classification model is constructed according to the crop identification key period and characteristic parameters determined by phenological information and spectral characteristics,so as to realize the monitoring research of crop planting structure change.The multi-source and multi-temporal remote sensing data can reflect the seasonal characteristics of different crops.Using the decision tree classification model constructed in this paper,the crop classification effect is better,with an overall accuracy of 87.54%.Based on the advantages of medium-high resolution and low-resolution remote sensing data,this paper uses the method of sample-based farmland information extraction and decision tree classification to monitor the planting area of main crops in Hailun City in successive years,and analyzes the change of planting pattern of main crops and its driving factors in the study area in the past six years.The planting patterns of corn,soybean and rice in Hailun City were extracted by the above method.Through the comparison and analysis with the agricultural statistical yearbook data,it is proved that the precision of the crop planting area obtained by remote sensing monitoring method is higher.According to the spatial distribution characteristics,the planting method of mixed cultivation of soybean and corn was the main method in Hailun City.The extraction results showed that the planting proportion of corn was about 58.3%,and the planting area decreased from 2014 to 2016,but the overall trend did not change much,but the area decreased significantly from 2016 to 2017.According to the extraction results,the planting proportion of soybeans was about 24.7% in 2014,and the planting area increased rapidly from 2015 to 2017,reaching about 37.6% in 2017.The rice planting proportion is the lowest,and the planting area is still declining in recent years.According to the change of crop planting area,the overall grain crop shows a downward trend,while the cash crop shows an obvious upward trend.From the perspective of natural factors,socio-economic factors and policy factors,this paper analyzes the influence of different driving factors on the crop planting pattern and the change of planting area of different crops.