The paper proposes a method for automatic classification of roads based on the use of a convolutional neural network Mask-R-CNN. The developed technique makes it possible to automate the task of categorizing roads, which is fundamental in the redistribution of traffic flows, since knowledge of the category of the road allows you to determine its maximum capacity. The article contains a description of the stages of training a neural network, as well as the results obtained when using it. The method of automatic road classification proposed in the paper showed good results both in classifying roads based on satellite images and in classifying roads based on photographs of road sections. When expanding the test set, the number of classes of recognized roads can be increased to match the categories of roads according to SP 34.13330.2021. In addition, this technique (in terms of segmenting objects in photographs) can be used to control the quality of the roadway.

Keywords: road categories, convolutional neural networks, satellite imagery, image segmentation, Mask R-CNN, image recognition, computer vision

The development of a methodology for the formation of an estimate of a construction object on the basis of its information model is considered, taking into account the state of development of software systems for BIM modeling and the peculiarities of regulatory regulation of the construction of construction projects in Russia. The factors that limit the possibility of drawing up estimate documentation for the information model have been identified. Taking into account the identified limitations, a set of operations necessary for the formation of an estimate for a construction object based on its information model is presented.

Keywords: BIM, 5D BIM-model, estimate documentation, civil and industrial construction

The article proposes a method of automatic recognition of the type of building for an environmental monitoring system. based on convolutional neural networks. To train the neural network, the Keras library was chosen, containing numerous implementations of the main components of neural networks, such as layers, target and transfer functions, optimizers, and many tools to simplify working with images and text. The processes of network implementation using the Google Colab cloud platform, the preparation of a training set, the training of a constructed neural network, and its testing during training are described. The result of this work is a convolutional neural network model, capable of determining with accuracy of the order of 90-92 percent what type of buildings is shown on the cartographic image, which allows us to automate this process and use it as a subsystem for the environmental monitoring system of atmospheric air.

Keywords: environmental monitoring system for air, building type recognition, convolutional neural networks, machine learning, computer vision

In article the algorithm of calculation of systems with unilateral constraints with the replacement reactions of the supports on the force variables. The calculation algorithm is based on the finite element method in the form of the classical mixed method. The algorithm of search of version of the working constructive scheme is based on the modified "physically obvious" algorithm of design constructively nonlinear systems. The paper describes the features of the resolution system of the algorithm proposed by the authors and provides a flowchart. A comparative analysis with the algorithm of calculation of systems with unilateral constraints, proposed by the authors earlier, as well as some others. The efficiency of the algorithm is tested on numerical examples.The advantages of the proposed algorithm are: The vector of results includes the reaction of the support, which allows you to make a decision to change the design scheme; the calculation takes into account, as an additional loading factor, the size of the displasement in the realized one-way support, which increases the accuracy of the calculation.

Keywords: structural mechanics, single-sided supports, structural nonlinearity, the mixed form of the finite element method