[Accuracy improvement of spectral classification of crop using microwave backscatter data]

Area of Science:

• Agricultural remote sensing
• Microwave remote sensing
• Multi-spectral imaging

Context:

• Accurate crop classification is crucial for agricultural management and food security.
• Satellite data fusion offers potential for enhanced land cover analysis.
• VV polarization microwave backscatter data provides structural information often complementary to spectral data.

Purpose:

• To investigate the use of VV polarization microwave backscatter data for improving crop spectral classification accuracy.
• To compare classification accuracies using fused multi-spectral and microwave data with single-source data.
• To evaluate the contribution of Envisat ASAR VV data to distinguishing crop types and field boundaries.

Summary:

• Fusion of HJ satellite multi-spectral data with Envisat ASAR VV polarization backscatter data was investigated for crop classification.
• The fused data leveraged spectral information from HJ and structural sensitivity from ASAR VV polarization.
• Results showed that fused data enlarged spectral differences, improving crop classification accuracy by up to 5% compared to HJ data alone.
• VV polarization data proved sensitive to non-agrarian areas and effectively distinguished field borders, enhancing classification.

Impact:

• The study demonstrates that fusing multi-spectral and VV polarization microwave data significantly enhances crop classification accuracy.
• This approach expands the utility of satellite data in agriculture, offering improved land cover characterization.
• The findings suggest broad potential for applying this data fusion technique in agricultural monitoring and management.