Answers are grounded in this paper and cite their sources.
The study demonstrates improved motion trajectory detection and image segmentation in ecological monitoring, suggesting better reliability in high noise environments.
In many ecological applications, e. g., for aerospace remote sensing of earth surfaces for the purpose of prevention of dangerous natural, ecological, and industrial catastrophes, it is necessary to perform the efficient detection of new moving objects, which could appear as a result of natural or other disasters. The method of detection and estimation of the objects’ motion as well as understanding of scenes on the basis of performance of the operation of segmentation of dynamical images is considered. The motion cue is used as a segmentation attribute of the object of interest. This method makes it possible to determine the motion trajectory, distance passed by a stealth object and its velocity as well as solve the problem of the reliable and trustworthy detection of the object of interest on the background of spatial interference. The article considers the possibility of the segmentation with the help of noise-correcting coding and of the expansion of the image signals in the basis of the Hadamard orthonormalized rectangular functions. The coding is performed on the basis of the use of finite cyclic group arithmetic and the concept of dyadic shift as an information parameter of a moving object. The main application of the method lies in the field of strong noise, real-time mode and limitations on computing and energy resources of observation. Under these conditions, the method allows to minimize time, computational, energy costs and detection errors in comparison with deep learning approaches. The algorithm can be used in the image classification process, for automation of identification of the objects on the shots obtained by sounding the earth surface on surveillance equipment such as small unmanned aerial vehicles.
No takes yet. Share an insight, caveat, or question.
Anatol Mitsiukhin (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: