Browse

This study explores the possibility of creating detailed maps of organic carbon content in the arable lands of northern Kazakhstan. The technology is based on filtering satellite images with convolutional neural networks and calculating the coefficients of the multi-temporal soil line using remote sensing data with a resolution of 30 meters. Calculating the coefficient of the multi-temporal soil line allowed us to create a map of the bare soil surface, highlighting areas of soil cover heterogeneity. The resulting ...