This study addresses a critical environmental issue in rice and wheat farming, where stubble burning significantly worsens air pollution in North India during winter. Farmers worldwide burn stubble after harvesting crops to quickly prepare fields for sowing the next crop. This practice contributes to dense smog in the Delhi-NCR region, exacerbated by winter winds carrying pollutants from neighboring states. The health and environmental consequences are severe. While sustainable alternatives such as composting and residue incorporation exist, stubble burning persists due to the high labor costs and time constraints involved in manual crop residue disposal. A well-processed and comprehensive dataset of crop residues is essential for developing advanced management technologies like AI and the IoRT to address this issue. However, the lack of such datasets online has hindered progress in this area. This paper presents a detailed image dataset of crop residues designed for use in image segmentation and classification models like Faster RCNN and Mask RCNN. The dataset, collected, preprocessed, and made available on Kaggle (at [ https://tinyurl.com/adhebla .]), is a valuable resource for researchers focusing on crop residue management. Furthermore, the study developed a Convolutional Neural Network (CNN) model with 98.96% accuracy to classify crop images into two categories: those with residue and those without. The model exhibited strong performance in residue detection. A user-friendly Graphical User Interface (GUI) was also integrated to enhance accessibility, allowing users to interact easily with the system.