研究目的
To monitor sugarcane harvest status using time series Sentinel-1 data.
研究成果
The proposed approach is capable of determining sugarcane patches and is able to discriminate harvested and non-harvested sugarcane areas with an overall accuracy of 82.17%.
研究不足
The producer accuracy of harvested sugarcane is comparatively lower due to partial harvesting in some fields.
1:Experimental Design and Method Selection:
The approach uses knowledge based classification and temporal profile for obtaining harvest status of crops.
2:Sample Selection and Data Sources:
Sentinel-1 SAR data, which is a C-band dual polarimetric data (VV and VH) is used in this study.
3:List of Experimental Equipment and Materials:
Sentinel-1 SAR data, GPS device for ground truth.
4:Experimental Procedures and Operational Workflow:
The data is first calibrated to obtain the backscattering coefficients, filtered to reduce speckle noise, terrain corrected, and then converted to dB scale. K-means clustering is used for segmenting the time series data.
5:Data Analysis Methods:
The average temporal profile of each patch is extracted, smoothened using Savitzky-Golay filter, and curve fitting approach is applied using smoothing spline.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容