Method of fragment based tracking of displacement of a large areal object in images

Abstract

Automation of technological processes using computer vision systems necessitates fragment based tracking of the displacement of a large areal object in images. Examples of such processes are monitoring mooring operations, loading and unloading operations, analyzing the environment from the vessel’s bridge, and tracking berthing. These processes are cyclical in the maritime transport industry and thus require methods for automatic data analysis in 24×7 mode. The paper proposes a method of fragment based tracking of the displacement of a large areal object in images in real time. The algorithm that underlies the method enables tracking the object fragment using convolution of pixel matrices and ensures data processing by frequency methods, including computational operations using Fourier images. The developed method is applicable for real-time tracking the object in images, even when it can be hardly distinguished from the background. The method of tracking the displacement of a large areal object proposed in the study can be used in software and hardware meters for longitudinal and vertical displacement, which are used in the maritime transport industry.