A method of spatial mapping and reclassification for high-spatial-resolution remote sensing image classification.

Area of Science:

• Remote Sensing
• Image Analysis
• Geospatial Information Science

Background:

• Accurate classification of high-spatial-resolution remote sensing images is crucial for various applications.
• Traditional pixel-based methods often struggle with spatial context, while object-based methods can be complex.
• Integrating different data types and classification strategies can potentially improve performance.

Purpose of the Study:

• To develop and evaluate a novel classification method for high-spatial-resolution remote sensing images.
• To effectively utilize both multispectral and panchromatic image information.
• To enhance classification accuracy by combining pixel-based and object-based techniques.

Main Methods:

• A four-step approach involving initial pixel-based classification (Support Vector Machine) of multispectral data.
• Segmentation of the panchromatic image using watershed segmentation.
• Spatial mapping of pixel classifications to segmented regions using an area dominant principle and thresholding.
• Reclassification of unclassified regions using spectral information (minimum distance to mean algorithm).

Main Results:

• The proposed method successfully integrates pixel-based and object-based classification strategies.
• It effectively leverages both multispectral and panchromatic information for improved classification.
• Experimental results demonstrate a significant improvement in classification accuracy for high-spatial-resolution remote sensing images.

Conclusions:

• The developed spatial mapping and reclassification strategy offers a robust method for remote sensing image classification.
• This hybrid approach enhances the utilization of multi-source remote sensing data.
• The method provides a valuable tool for accurate land cover mapping and analysis.