The tobacco business continues to experience difficulties adhering to regulations, particularly regarding the packaging of cigarettes. It can be computationally demanding, needing strong hardware for real-time applications, and it might have trouble with severely damaged or concealed packaging. We present a new technique for the analysis of cigarette packaging in this paper named Pelican-driven Tuned Convolution Kernel ResNet (P-TCKR). Pelican optimization improves the performance of the convolutional kernel in the ResNet framework, enabling more precise and effective quality evaluations of cigarette packaging. Three primary classifications were represented by the varied range ofcigarette package images in our dataset. We used a bilateral filter in the data pre-processing step to improve the quality of the input images and lower noise. The suggested P-TCKR framework is tested on the Python platform and examined using F1-score (91.50%), accuracy (91.70%), recall (92.60%) and precision (92%) measurements. P-TCKR is a major step forward in the development of effective and dependable quality control solutions for the analysis of cigarette packaging.