UAV Video Processing for Traffic Surveillence with Enhanced Vehicle Detection

Remotely sensed data using aerial video is gaining vast popularity in traffic surveillance. Traditional monitoring devices are usually kept at fixed locations to achieve a fixed surveillance coverage range. Unmanned aerial vehicles (UAVs) are receiving much attention from researchers in traffic monitoring due to their low cost, high flexibility, and wide view range. Unlike stationary surveillance, the camera platform of UAVs is in constant motion and makes it difficult to process for data extraction. To address this problem, a novel framework for traffic flow parameter estimation with enhanced vehicle detection from aerial videos is proposed. The frames initially undergo Adaptive Gamma Correction for contrast enhancement. The proposed method incorporates steps that make use of the Kanade-Lucas-Tomasi (KLT) tracker, Support Vector Machine (SVM), and connected graphs. Interest-point-based motion analysis is done using KLT tracker. The SVM classification is used to identify vehicles from other moving objects. Contourlet transform coefficients and Gray Level Co-occurrence matrix (GLCM) are among the seven features extracted from the frames. Finally the number of vehicles, average speed and densities of bi-directional flow are found. The method shows high detection rate and low false positive alarms.