Vehicle Detection Network using CNN and instance Segmentation

:- These days, when an accident happens, people either hide from the emergency services or cause a lot of disturbance while reporting it, or the accident goes unreported and by the time help arrives, it’s too late. A comprehensive system has been developed to actively detect all types of traffic accidents and notify the appropriate parties. In the event of an accident, this includes the closest police station, hospitals, general ambulances, the registered owner of the vehicle involved in the accident, and their emergency contacts. In the event of a hit-and-run, the police can obtain the vehicle number of the accused vehicle. Numerous applications exist for vehicle detection in aerial pictures, and most vehicle detection techniques employ the bounding-box approach for localization. Using a private dataset, a convolutional neural network is proposed in this letter to detect cars. Based on experimental data, CNN performs better, with an accuracy rate of 94.54%.