Satellite imagery recognition is a vital application of computer vision, with numerous potential use cases in fields such as disaster management, land use, and urban planning. Convolutional Neural Networks (CNNs) have attained significant success in this field. This study aims to determine the optimal quantity of input data required for training CNNs for satellite imagery recognition. To achieve this objective, a series of rigorous experiments are conducted to examine the impact of input data quantity on various performance metrics, including accuracy, convergence, and model generalization. The findings reveal that an increase in input data quantity generally improves model performance up to a saturation point, beyond which further increases do not lead to significant improvements and may even lead to overfitting. This study proposes several strategies for identifying this saturation point and mitigating the effects of overfitting. The insights gained from this research can aid practitioners in designing more efficient and effective CNN models for satellite imagery recognition, a critical and challenging task in computer vision.