ImageNet-Architect Model for Accurate Cloud Sky and Sun Image Segmentation

Cloud, sky, and sun segmentation is a crucial task for various applications, including weather analysis, environmental monitoring, and image enhancement. Motivated by the need for accurate and efficient segmentation methods, the proposed ImageNet-Architect model, a robust encoder-decoder architecture designed to segment images into three distinct classes: clouds, sky, and sun. The aim of this study is to establish a reliable segmentation standard that can accurately classify each pixel in an image into the respective categories, thereby facilitating more precise analysis and interpretation of sky conditions. The model employs convolutional layers, batch normalization, ReLU activations, and skip connections to capture high-resolution features and maintain spatial information. The ImageNet-Architect model on a dataset comprising 512 images and their corresponding ground truth masks. The model was trained using data augmentation techniques to enhance its robustness to varying conditions. The results demonstrate the model's high accuracy, with Intersection over Union (IoU) scores of 0.88 for clouds, 0.91 for sky, and 0.87 for sun. Visual comparisons between the segmented outputs and ground truth masks indicate that the model can effectively delineate the boundaries of clouds, sky, and sun under diverse conditions. The total parameters of the model are 31, 055, 297, of which 31, 043, 521 are trainable, reflecting the model's capacity for detailed feature learning.