CAJ | 학술논문

GF-1号卫星是中国2013年4月26日发射的一颗高分辨率遥感卫星，为解决该新型卫星数据在农作物对地抽样遥感调查中的应用技术方法问题，该文针对GF-1号卫星数据的特点，研究了基于GF-1号卫星16 m WFV传感器和2 m/8 m PMS传感器卫星数据的农作物种植面积遥感抽样调查方法。根据研究区物候历，选择农作物识别关键期的16 m WFV传感器数据进行多时相农作物种植面积的中分辨率遥感提取；在中分辨率农作物面积遥感分类图基础上，计算研究区域的MORAN I指数，确定格网抽样单元的大小，进行多目标农作物的MPPS（multivariate probability proportional to size）抽样；对抽样单元采用2 m/8 m PMS传感器卫星数据进行高分辨率农作物面积制图；最后根据MPPS抽样方法进行总体农作物种植面积的推断，并计算CV值，评价抽样精度。以江苏省东台市为研究区对GF-1号卫星数据进行了应用研究。研究结果表明，GF-1号卫星数据完全可以应用于县级农作物种植面积的提取，农作物种植面积提取精度优于90%。
GF-1호위성시중국2013년4월26일발사적일과고분변솔요감위성，위해결해신형위성수거재농작물대지추양요감조사중적응용기술방법문제，해문침대GF-1호위성수거적특점，연구료기우GF-1호위성16 m WFV전감기화2 m/8 m PMS전감기위성수거적농작물충식면적요감추양조사방법。근거연구구물후력，선택농작물식별관건기적16 m WFV전감기수거진행다시상농작물충식면적적중분변솔요감제취；재중분변솔농작물면적요감분류도기출상，계산연구구역적MORAN I지수，학정격망추양단원적대소，진행다목표농작물적MPPS（multivariate probability proportional to size）추양；대추양단원채용2 m/8 m PMS전감기위성수거진행고분변솔농작물면적제도；최후근거MPPS추양방법진행총체농작물충식면적적추단，병계산CV치，평개추양정도。이강소성동태시위연구구대GF-1호위성수거진행료응용연구。연구결과표명，GF-1호위성수거완전가이응용우현급농작물충식면적적제취，농작물충식면적제취정도우우90%。
The Chinese GF-1 satellite is a new high spatial resolution satellite launched on April 26, 2013. It was equipped with two types of sensors. One is the wide field view sensor (WFV sensor); the other is the panchromatic and multispectral sensor (PMS sensor). The WFV sensor can acquire multispectral image in blue, green, red, and near-infrared bands with 16 meters spatial resolution and 4 days temporal resolution. The PMS sensor can acquire a panchromatic and multispectral image with 41 days temporal resolution. The spatial resolution of a panchromatic image acquired by the PMS sensor is 2 meters, while the spatial resolution of a multispectral image acquired by the PMS sensor in blue, green, red, and near-infrared bands is 8 meters. According to the characteristics of a GF-1 satellite image, a method for mapping crops using remote sensing and sampling technology was proposed. There are four kinds of summer harvest crops in our study area of Dongtai county. There are winter wheat, barley, rapeseed and vegetables. According to the crop’s phenology calendar information of this study area, there are three key periods for the identification of crops. In later March, winter wheat and barley are in the growing season, while rapeseed is in the flowering period. In early April, barley and rapeseed are in the flowering period, while winter wheat is in the growing season. In early and middle May, winter wheat is in the flowering stage, while canola and barley are at maturity. So the 16 meters resolution WFV sensor data acquired in those periods were used to classify those crops. First, that data was preprocessed for ortho rectification, geometric correction, and atmospheric correction. Then multi-days NDVI were calculated and was used to generate a false color composite image. In the false-color composite image, we found that those kinds of crops exhibited distinctly different colors. Vegetables were yellow, canola was light red, grain crop including winter wheat and barley is dark red. So those crops can be easily classified using the maximum likelihood method. Then we converted the classification map to vector files and calculated a MORAN index in this study area using ARCGIS. The MORAN index in this study area was 0.78, and the distance threshold was 5480 meters. The resolution of WFV data was 16meters, so 5440meters (340pixels of WFV) was set to the size of the sampling units. Then 10 sampling units were selected using the MPPS method based on the proportional information of each crop for each unit which was calculated using the block statistics function of ARCGIS. After that, the high spatial PMS sensor data of those 10 sampling units were resized out. In the high spatial false-color composite PMS sensor image, we found that those crops also had different colors. Rapeseed was light red, wheat was dark red, and vegetables were gray. So the crops in each of the sampling units were classified with the maximum likelihood method using the field investigation data as the training sample. Finally, according to the MPPS method, the total areas of each crop in the study area and CV were calculated. The area of crops, rapeseed, and Vegetables of 2014 in Dongtai using temporal WFV data was 582.74km2, 226.7873km2, and 271.288km2. Compared with the data published by the Dongtai County Farm Bureau, the errors of crops, rapeseed, and vegetables using temporal WFV data were -9.9%, -2.8% and -15.2%. The area of crops, rapeseed, and vegetables of 2014 in Dongtai County using remote sensing and sampling methods was 638.6318km2, 244.8km2, and 322.9601km2. Compared with the data published by the Dongtai County Farm Bureau, the errors of crops, rapeseed, and vegetables using remote sensing and sampling methods were 3%, 5%, and 1%. Those results showed that this method could classify crops, rapeseed, and vegetable areas effectively. High mapping precision of 90% was acquired.