Nowadays vehicles are being used a lot and different types of vehicles come at a time. Many vehicles running at a time leads to collisions and many people are dying in accidents. Pollution and high speed vehicles are running in cities and due to this more than 1.25 million people are dying every year as a result of road traffic accidents globally. Even after that no such technology has been available which gives recommendations in real time to prevent accidents. But my system is providing a good solution to this problem which detects the collision of traffic vehicles and provides collision free route. The aim of my research paper is to solve this challenge by developing an optimal recommendation model on top of a geographic information system in which collision free route is provided using advanced planning techniques. Urban transportation systems face significant challenges due to traffic congestion and accident-prone zones, leading to increased delays, environmental pollution, and safety concerns. Traditional methods often fall short in delivering real-time, congestionminimized route recommendations and typically do not integrate accident zone detection with traffic management. To overcome these limitations, this study introduces GINSER—a Geographic Information System (GIS)-based optimal route recommendation model powered by advanced deep learning techniques. The main goals of the GINSER framework include identifying accident-prone areas, classifying traffic congestion levels, and recommending efficient travel routes using GIS analysis. CCTV footage is used as the primary input, with image quality enhanced through an Adaptive Gaussian Bilateral Filter (AGBF) for noise reduction. A Faster R-CNN model is employed for object detection and localization within high-risk zones, while Particle Swarm Optimization (PSO) fine-tunes hyper parameters to boost detection accuracy. To classify traffic congestion into four categories—low, moderate, high, and congestion-free—a CNN-BiGRU architecture is utilized. GIS further processes spatial and traffic data to suggest the most efficient, congestion-free routes. The model’s performance was evaluated using metrics such as F1 score, accuracy, precision, recall, and specificity, confirming the effectiveness of the proposed GINSER system. The noise-free images using AGBF effectively enhances image quality by reducing noise leading to improved classification accuracy. PSO is utilized for hyperparameter tuning achieving a high accuracy of 98.24%. The GINSER model achieved a classification accuracy of 99.16%. The GINSER improved overall accuracy by 4.90%, 6.83%, 5.23%, and 0.83% better than TSANet, TCEVis, Ising-traffic, and AID, respectively. Its ability to detect accident prone zones classify congestion levels and recommend optimal routes ensures safer and more efficient mobility