This research investigates the use of deep learning-based computer vision techniques for monitoring road geometry changes to support urban planning and infrastructure management. Traditional road monitoring methods are often limited by time and cost, which necessitates an automated system capable of detecting and analyzing structural changes using video and image data. The proposed system consists of two custom-trained models: one for detecting Right of Way (ROW) and classifying surrounding land use types (residential, industrial, water bodies) around road boundaries, and another for identifying roadside vegetation. These models provide insights into unauthorized encroachments, vegetation distribution, and areas that need environmental improvements.  The approach involves processing input data through deep learning algorithms, converting video frames and images into quantitative insights that reveal structural changes over time. The models are trained using diverse urban road datasets and achieve reliable accuracy in detecting both road boundaries and vegetation. This project presents the system’s design, implementation, and performance, highlighting the potential of AIdriven solutions in transforming road infrastructure management