Smart driving: Developing image processing and control approaches for safer roads and enhanced mobility

Currently, with the increase in the number of vehicles in modern cities, the parking problem and accidents have emerged as significant issues. The search for parking spaces and driver distractions in public areas contribute to traffic congestion, environmental pollution and time loss. The aim of this research is to contribute to parking management and safely driving solutions by using image processing and shortest path planning machine learning algorithms. Additionally, it focuses on advancing the Smart Transportation Systems (STS) through implementing the Aggregate Channel Features (ACF) algorithm to detect vehicles and determine their positions where the occupancy map of parking areas was determined by using determined locations. Moreover, these vehicle positions were used in the analysis of driver behavior for recognizing driver intentions. In the parking management algorithm, the Breadth-First Search (BFS) algorithm was modified to determine the shortest path to the target parking slot. This algorithm guides the newly entered vehicle from the entrance gate to the target parking area. In the driver intention algorithm, which is developed to prevent accidents, neighboring vehicle positions and their driving movements were analyzed and labeled as "calm" and "beware". It has performed effectively in various testing scenarios, where the vehicles were correctly detected and labeled correctly according to their movements. In addition, the parking management algorithm was highly accurate in detecting the vehicles within the parking areas. It effectively guided the drivers to their allotted parking slots where it resulted in a 61.8% reduction in searching time. The results confirm that this research made a significant contribution to sustainability by improving parking management and driving safety while also reducing emissions and resource waste.