Decision making algorithms in the problem of object selection in images of ground scenes

Detection and localization of objects in images of 3D ground scenes

“…Thus it is known a priori that the desired object belongs to the image (only the coordinates of the target point location in the CI are unknown). This paper logically continues paper [3] specified above in the application not only the information about image contours and references of objects, but also infor mation about angular orientation of image edges and references of these objects and angular orientation of boundaries of patterns and references of salient terrain fragments.…”

“…However distinctions between these problems take place in all stages of decision making: problem statement, methods of specifying prior information, formulated statistical hypotheses, decision rules applied, and other aspects of solving these problems; in detail the similarity and distinctions of the selection problem and the recognition problem were presented in [4]. The solution of the problem of the problem of detection and localization of a given object in an image of a ground 3D scene [3], when the conditions of its observation (errors in the navigation system of the PV, errors in orientation of the optical axis of the onboard sensor of the CI to the calculated point, errors of prior information of the observation scene, in the parameters of the onboard sensor of the CI, etc.) are such that the image of the given object with the probability equal to one, is present in the CI is an example of practical application of this approach.…”

“…Note that frequently for solving par ticular technical problems both known (obtained and applied in solving other problems) and heuristic algorithms are used without sufficient substantiation. In order to solve the problems formulated above, the authors proposed and described in a number of papers, e.g., in [1][2][3], algorithms of structural-statistical type, which in our opinion take into account the specific features of the posed problem for the scenes of the specified types most completely. The main stages of implementation of the specified algorithms are as follows:…”

“…The analysis of capabilities of the algorithmic implementation of these stages (especially the choice of features for generating the reference description of images of the observation scene and decision making) has shown that many image processing and analysis problems important in applications can be posed and solved as problems of selection in the CI patterns of given objects [3][4][5] rather than as recognition prob lems. It is important to note that this approach makes it possible to simplify substantially both the solution of the problems specified above and the design of the whole technical system in formalization of the applied problem itself, its decomposition, reduction requirements to the amount of prior information, etc.…”

A method for localizing a reference object in a current image with several bright objects

“…One of the stages in selecting a given object in the image of a complex three-dimensional terrestrial scene obtained by the onboard sensor of an unmanned aerial vehicle is considered in [9]. However, the algorithm that implements a set of procedures for processing and analyzing images as well as deciding on the coordinates of the location of a given object in the CI is designed without taking into account the content of the image scene.…”