Distance estimation between moving objects using monocular camera

Distance estimation between two or more objects is a crucial task in the computer vision research area. Moreover, in the era of COVID-19, it becomes an urgent issue as it can enable social distance preserved. Distance estimation could be done using stereo vision (stereoscopic photogammetry) but requires more complexity. In this paper, researchers show distance estimation is possible using only monocular vision. We propose a deep-learning based method, Mobilenet Single Shot Detector (MSSD), combined with Camera Calibration to detect objects and estimate the distance between them in the setting of monocular vision. To verify the robustness of the proposed method, we created a dataset video using a monocular camera. The experimental results showed the performance of the proposed method could estimate the distance properly using the recorded dataset.