Comparative Study between Two Methods of Crop Classification in the Irrigated Area of Sidi Bennour

Abstract

Knowing crop types, surface area, and spatial distribution is essential for monitoring and evaluating their vegetative states. Indeed, this information is crucial for management decision-making in the agricultural sector, especially in irrigated sectors, such as the Sidi Bennour case, where agricultural activity is intensive. This area is part of the Doukkala irrigated perimeter known, in Morocco, for its importance in agricultural production. Remote sensing has become essential for monitoring the vegetation state and crop mapping. Our study's main objective is crop mapping using earth observation data with three spatial resolutions: 10, 15, and 30 m (Sentinel-2, pancharpened Landsat-8 image, and the original Landsat-8 image, respectively). Two classification methods, Support Vector Machine (SVM) and Maximum Likelihood (ML), have been applied to discriminate the different crop types. The SVM method gave the best results for all three spatial resolutions. Also, pansharpening has improved the classification for the Landsat-8 image.