This work is devoted to the problem of detecting and studying archaeological sites based on methods of machine learning, mapping, and geophysics. The article discusses the methodology for mapping and surveying archaeological objects in the form of 2D and 3D models, and proposes machine learning methods based on convolutional neural networks for the detection and classification of archaeological objects.

In this paper a system for the detection and research of archaeological sites using on machine learning methods, mapping methods and geophysics methods is presented. The development of an information system for remote archaeological research allows, on the one hand, to solve the most important problem of archaeological science, related to the preservation of cultural heritage, and on the other hand, to reduce the duration of the archaeological examination procedure conducted before the industrial development of the territory. In this paper the method for registering 3D data based on ICP (Iterative Closest Points) based on fusion of visual and semantic characteristics of an archaeological site for a class of orthogonal transformations was proposed. Computer modeling of the proposed 3d registration method was carried out using a collection of data from archaeological sites of the Bronze Age of the Southern Trans-Urals. When performing computer modeling, methods for solving the variational ICP problem were considered, using various options for minimizing the functional: the point-to-point metric with and without extrapolation, using the point-to-plane metric. A comparative analysis with various options for minimizing the functional was carried out, the results of computer simulation were presented and discussed.

A closed-form solution is proposed for the problem of minimizing a functional consisting of two terms measuring mean-square distances for visually associated characteristic points on an image and mean-square distances for point clouds in terms of a point-to-plane metric. An accurate method for reconstructing three-dimensional dynamic environment is presented, and the properties of closed-form solutions are described. The proposed approach improves the accuracy and convergence of reconstruction methods for complex and large-scale scenes.

New combined algorithm for simultaneous navigation and map construction is developed using visual characteristics and depth information to compare images, register 3D-point clouds, and build global sequential 3D-maps of the surrounding space. The performance and computational complexity of the proposed RGB-D SLAM algorithm are presented and discussed with reference and real data. The results can be applied in real-time tracking of objects, in non-cooperative remote observation, and semantic mapping of mobile robot navigation problems.

In this paper a new method will be proposed of determining the dynamic position of a robot in a relative coordinate system based on Kalman filtering, on a history of camera positions and on the robot's movements, on symbolic (semantic) tags. In order to track the robot's reiterated passage of one and the same place, it is necessary to carry out, at each step, the matching of the robot's position and state with the previous steps (the problem of << loop closure >> - a loop closure and global optimization step). In the event of data coincidence, it is necessary to carry out adjusting the movement and refining a three-dimensional map of the environment. One of the known solutions of this problem will be taken as a basis and improved in the present work based on the algorithm of << the basket of words >>. We evaluate the RGB-D Loop-closure detection in indoor environments of Chelyabinsk State University.

Nowadays many algorithms for mobile robot mapping in indoor environments have been created. In this work we use a Kinect 2.0 camera, a visible range cameras Beward B2720 and an infrared camera Flir Tau 2 for building 3D dense maps of indoor environments. We present the RGB-D Mapping and a new fusion algorithm combining visual features and depth information for matching images, aligning of 3D point clouds, a “loop-closure” detection, pose graph optimization to build global consistent 3D maps. Such 3D maps of environments have various applications in robot navigation, real-time tracking, non-cooperative remote surveillance, face recognition, semantic mapping. The performance and computational complexity of the proposed RGB-D Mapping algorithm in real indoor environments is presented and discussed.

A new method will be developed in the present work of the detection of a robot's position in a relative coordinate system based on a history of camera positions and the robot's movement, symbolic tags and on combining obtained three-dimensional depth maps that account for accuracy of their superimposition and geometric relationships between various images of the same scene. It is expected that this approach will enable one to develop a fast and accurate algorithm for localization in unknown dynamic environment.

In this work we present an algorithm of fusing thermal infrared and visible imagery to identify persons. The proposed face recognition method contains several components. In particular this is rigid body image registration. The rigid registration is achieved by a modified variant of the iterative closest point (ICP) algorithm. We consider an affine transformation in three-dimensional space that preserves the angles between the lines. An algorithm of matching is inspirited by the recent results of neurophysiology of vision. Also we consider the ICP minimizing error metric stage for the case of an arbitrary affine transformation. Our face recognition algorithm also uses the localized-contouring algorithms to segment the subject's face; thermal matching based on partial least squares discriminant analysis. Thermal imagery face recognition methods are advantageous when there is no control over illumination or for detecting disguised faces. The proposed algorithm leads to good matching accuracies for different person recognition scenarios (near infrared, far infrared, thermal infrared, viewed sketch). The performance of the proposed face recognition algorithm in real indoor environments is presented and discussed.

A face recognition method based on a matching algorithm with recursive calculation of oriented gradient histograms for several circular sliding windows and a pyramidal image decomposition is proposed. The algorithm produces good results for geometrically distorted and scaled images.

In this article we present a method of extracting factual data (objects, their attributes and relationships) from natural language texts using ontological knowledge base model. Subject area is represented through ontological model extended by fuzzy links between objects.