![]() However, the model cannot be readily extended to produce maps in other years due to its limited generalization ability. 14 generated the distributions of maize in Heilongjiang province from 2015 to 2016 by applying a random forest model on Landsat 8 images. Although the spatial resolution of Landsat images is lower than that of the Sentinel images, Landsat can provide long-term earth observation over 50 years, which has become the indispensable foundation for long-term and large-scale land use analysis.Ī notable example of maize mapping studies using Landsat is the CropScape, displaying data from the United States Department of Agriculture (USDA) Cropland Data Layer (CDL), which adopted machine learning methods to generate maps with more than a hundred crops including maize, across United States 13. To address the challenge, a plausible solution is to generate maize maps with high-spatial resolution satellites such as Sentinel and Landsat. Although daily MODIS data can provide continuous phenological information of crops with a high temporal resolution, its coarse spatial resolution cannot accurately capture the crop distributions in China dominated by small-scale farmlands. Traditionally, most maize maps are generated by performing crop classifications with MODIS satellites in the 500 m to 1 km coarse spatial resolution range. Over the last decades, several methods have been proposed to map maize with optical remote sensing 8, 9, 10, 11, 12. Therefore, mapping the one-season maize cropland in China will be of great significance to ensure food and energy security 6, 7. Nonetheless, due to rapid urbanization and industrialization, global temperature rises, soybean rejuvenation, as well as changing precipitation patterns, the one-season cropland of China, which produce maize with high nutritional value, are at risk 4, 5. According to the statistical corporate database released by the food and agriculture organization, China has become the second biggest maize-producing country, contributing around 23% of the production of global maize 3, which is crucial for both food and energy security. Due to its rich starch content and comparatively easy conversion to ethanol, maize is not only important for global food security but also a popular feedstock for biofuel production all over the world. The maize-cultivated areas derived from the maps are highly consistent with the data recorded by statistical yearbooks ( R 2 = 0.85 on average), which indicates that the produced maps are reliable to facilitate the research on food and energy security.Īs one of the most widely distributed staple food, maize is grown in at least 164 countries and produces more than 35 percent of the world’s food 1, 2. With the generalization capability, the proposed method produces maize cropland maps with a resolution of 30 m from 2013 to 2021 in the one-season planting areas of China. In this paper, we collect 75,657 samples based on field surveys and propose a deep learning-based method according to the phenology information of maize. Nonetheless, due to the lack of survey data related to planting types, long-term and fine-grained maize cropland maps in China dominated by small-scale farmlands are still unavailable. ![]() ![]() Quantifying the area changes of maize cropland is crucial for both food and energy security. As the major maize-cultivated areas, the one-season cropland of China is increasingly threatened by rapid urbanization and soybean rejuvenation. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |