Real time traffic video analytics using drones

Overview

Unmanned Aerial Vehicles (UAVs) are becoming an attractive solution for road traffic monitoring because of their mobility, low cost, and broad view range. The clients approached us with a need to analyze traffic videos and image on roadways taken using drones and CCTV videos. They wanted to leverage the CCTV video recordings and generate insightful analytics about the traffic on residential and commercial properties. The clients wanted the system to predict and forecast alternate routes to avoid traffic congestion on roads. An interesting requirement was to detect anamolies on the vehicle movement to prevent accidents and untoward road rage incidents.

Challenge

We developed a Deep Neural Network model to recognise six categories of vehicles namely cars, motorcycles, buses, trucks, bicycles, and pedestrians. We first collected the CCTV footage and generated a customized data set containing 2000 images for each class of vehicle. We then trained a Deep Neural Network to classify each class of a vehicle and localize them on a video frame. Since this was a video processing application, the performance was a key metric in addition to the accuracy of the vehicle finder.

Solution

We developed a Deep Neural Network model to recognise six categories of vehicles namely cars, motorcycles, buses, trucks, bicycles, and pedestrians. We first collected the CCTV footage and generated a customized data set containing 2000 images for each class of vehicle. We then trained a Deep Neural Network to classify each class of a vehicle and localize them on a video frame. Since this was a video processing application, the performance was a key metric in addition to the accuracy of the vehicle finder.

Impact

The designed traffic analytics system was reliable and performed with an accuracy of 87% . We were able to extract useful information about traffic patterns matching different times, days and even weather condition. The collection of data consisting of images and videos provided useful insights for their rerouting plans and strategies and we were able to imrpove their road congestion problems by almost 25%. The vehicle monitoring system was able to detect accidents and road rage scenarios with great accuracy and accidents reduced by 30%.

Technology stack