 The study developed an integrated fuzzy object based image analysis and deep learning approach, PhobiaDL, for monitoring land use cover changes and compared it to three machine learning algorithms. The PhobiaDL approach used fuzzy membership functions and deep convolutional neural networks, DCNNs, to derive object features and attributes, while the fuzzy synthetic evaluation and Dempster Schaefer theory, FSEDST, apply to validate and carry out spatial uncertainties. The results showed that the PhobiaDL approach outperformed the other approaches and improved the strength and robustness of OBIA decision rules, while the FSEDST approach improved the spatial accuracy of object-based classification maps. The study supports lake restoration initiatives by decision makers and authorities in applications such as drought mitigation, land use management and precision agriculture programs. This article was authored by Baxia Faisizadeh, Kivan Mohamedzadeh Ladjajeh, Tobia Lakes, and others.