Optical Tracking of Floating Shipping Containers in a High-Velocity Flow

Abstract

Debris in high-volume, high velocity flows, such as tsunamis, storm surges, and dam breaks, cause widespread damage to coastal communities. Evaluating the motion of the debris within the flow is difficult due to the variety of variables that can affect the motion, therefore techniques must be developed to quickly and accurately track the motion of the debris within the flow. This paper presents an optical tracking technique to track the motion of debris over a horizontal bottom under laboratory conditions. The debris consisted of scaled-down shipping containers, made of polyethylene (PE), placed on a dry, raised, flat section of the basin floor. Container movement was measured using two overhead cameras and a novel object tracking and detection algorithm implemented in MATLAB. The algorithm was developed to be able to quickly and accurately track uniform containers throughout an experimental run. The algorithm combines image-processing techniques such as Color Thresholding and Blob Analysis with tracking methods such as the Kalman Filter and Hungarian (Munkres) Algorithm. The algorithm was evaluated by comparing the results to those from a manual tracking method. For experiments with one, three, and six containers, the algorithm showed good agreement with the results of the manual method. However, with nine containers, the increase in the amount of container—container collisions and the agglomeration of the containers resulted in poor results compared to the manual method. While the algorithm is only evaluated here to track the motion of the shipping containers, there is potential for its use in many other fields of hydraulic and coastal engineering.