Scholarly record

CROP MAPPING USING HYPERSPECTRAL DATA AND TECHNOLOGIES - A COMPARISON BETWEEN DIFFERENT SUPERVISED SEGMENTATION ALGORITHMS

Mohamad M. Awad

Abstract

Hyperspectral images and field spectroradiometer with high spectral resolution can improve substantially crop mapping by reducing similarities between different crop types which has similar ecological conditions. This is done by recording fine details of the crop interaction with sunlight. This paper deploys crop spectral signatures database interactive tool for the major crops in the Eastern Mediterranean Basin. The collection of spectral signatures of crops is performed during the growth stage of the crops. The database includes several physical and chemical parameters for crops, resampled spectral signatures for a specific multispectral or hyperspectral satellites. Combining the hyperspectral data with an efficient segmentation algorithm can increase the accuracy of the final crop map. To prove this idea, major crops such as ?winter wheat? and ?spring potato? are mapped using the hyperspectral data which includes spectral signatures database and CHRIS-Proba satellite images. The images are segmented using different supervised algorithms. The evaluation of the results showed that using the database interactive tool with Spectral Angle Mapper (SAM) algorithm increased the accuracy significantly.

Publication details

Back to publication list