Maps analysis of Najran City, Saudi Arabia to enhance agricultural development using hybrid system of ANN and multi-CNN models

Abstract Agriculture in Najran City plays a crucial role in supporting the local economy. Nevertheless, the region faces sustainability, resource management, and environmental degradation challenges. Descriptive and somewhat manual techniques for assessing land-use scenarios and classifying agricultural land changes can be tedious and prone to human error, and they remain inefficient when analyzing rapidly changing terrains. AI techniques offer promising solutions to overcoming such shortcomings, enabling automated, accurate, and scalable analyses of highly intricate satellite imagery datasets. This study proposes a hybrid artificial intelligence-based topographic analysis model to improve agricultural development in Najran. Spatial features extraction from the images of those two models through spatial information fusion would, in the post-fusion, mean an increase in redundancy and irrelevant features and an increase in computational complexity. Therefore, the next step was using a PCA algorithm to remove unessential features before feeding essential features into the artificial neural network (ANN) algorithm. The resulting important features are then classified using an ANN. The system was then implemented using Landsat 8 satellite imagery for 2013–2023. The results indicated that the developed hybrid EfficientNetB7–ShuffleNet with ANN system achieved an accuracy of 97.11% in 2013 and 97.01% in 2023. It also achieved a Recall rate of 97.66 and 97.54% during the period 2023 and 2023, respectively, and an F 1-score of 96.77 and 97.53% during the period 2023 and 2023.