GEOMEMBRANE BASINS DETECTION BASED ON SATELLITE HIGH-RESOLUTION IMAGERY USING DEEP LEARNING ALGORITHMS

Abstract. Agriculture is a very important economic sector in Morocco, which requires set of tools to improve agricultural production. Among these tools, the use geomembrane basins. The latter great importance smart farming planning, management practices or even livestock use. In this context, study evaluates recognition and classification basin using remote sensing satellite images; based on Yolov3 deep learning neural network. This paper first adjusts network model make it suitable for detecting small targets images, then uses k-means algorithm calculate grid size Yolo basins, yolov3 train data that makes up imagery. detection basins obtained by test phase. Finally, adapted image validation Through research analysis experimental results, can be seen method effectively detects images ensures high accuracy gave us an 75%.