Speed Estimation of Multiple Moving Objects from a Moving UAV Platform

Abstract

Speed detection of a moving object using an optical camera has always been an important subject to study in computer vision. This is one of the key components to address in many application areas, such as transportation systems, military and naval applications, and robotics. In this study, we implemented a speed detection system for multiple moving objects on the ground from a moving platform in the air. A detect-and-track approach is used for primary tracking of the objects. Faster R-CNN (region-based convolutional neural network) is applied to detect the objects, and a discriminative correlation filter with CSRT (channel and spatial reliability tracking) is used for tracking. Feature-based image alignment (FBIA) is done for each frame to get the proper object location. In addition, SSIM (structural similarity index measurement) is performed to check how similar the current frame is with respect to the object detection frame. This measurement is necessary because the platform is moving, and new objects may be captured in a new frame. We achieved a speed accuracy of 96.80% with our framework with respect to the real speed of the objects.

Vision‐based vehicle speed estimation: A survey

“…It can also be performed either manually [20, 38, 58] or automatically [12, 14, 18, 19, 23, 33, 78, 85, 88, 101, 108]. In some limited cases, camera calibration is somehow neglected [13, 15, 16, 21, 22, 24–26, 30, 32, 34, 39–41, 43, 44, 46, 48–53, 55, 56, 60–65, 69, 71, 73, 81, 84, 95, 103–107, 111, 113, 118, 120, 124].…”

Vehicle Detection In Remote Sensing Images

“…However, the data collected from different sensors mounted in the vehicles. Biswas et al [23], used speed estimation for moving object detection based on UAV system. Faster R-CNN is applied to detect the objects, and a discriminative correlation filter with CSRT is used for tracking.…”

Road Traffic Monitoring from UAV Images Using Deep Learning Networks

“…There have also been various attempts to automatically measure traffic from road videos taken by UAVs [11][12][13][14]. Most of these methods consist of object detection techniques [15] for capturing vehicles in images and object tracking techniques for identifying the movements of detected vehicles, and the speed of vehicles are calculated at the end.…”

“…This approach requires a pre-definition of a structure for each location; therefore, images without known structures cannot be utilized. Later, the image scale is calculated by comparing the average sizes of vehicles in the images and pre-measured and averaged actual vehicle size in [12,20]. Although these methods have somewhat resolved the restrictions associated with a UAV's flight area, the calculated image scale is not accurate because the size of vehicle varies depend on the types of the vehicles.…”