Study of the quality construction of photogrammetric models obtained in low-light conditions

Abstract

Relevance of the research topic. The paper shows main results of environmental parameters that affect the quality of 3D-models obtained by the photogrammetric method of surveying. Research objective. The article presents: the fundamental possibility of using a small-sized photogrammetric device for underground surveying and the parameters that affect the result of shooting in a confined low-light space. Research methods. Experiments included a set of independent serial photographs of a test site. To increase the results of experiments, object with a simple geometric shape have been used as test objects for extrapolating of the results of the survey to products of standard metal products. In these conditions, imitating abstract underground mine workings: with the lack of lighting, confined spaces, a small number of characteristic points in the background. The texture of the subject of photogrammetric survey and common geometric parameters of photo base accepted as studied factors. Various images were projected onto the test scene to form an artificial texture. Results and its applications. The experimental results confirm the possibility of using the proposed methods for identifying engineering communications nodes, oil and gas production facilities, elements of lifting machines and conductors in vertical mining devices. In addition, the authors performed an idea of improving the quality of photogrammetric models by projecting temporary artificial textures on the subject. The results of the work can be used for the improving of the methods of surveying, using digital photogrammetry technology Findings. It is possible to imrove the quality of digital photogrammetric models of engineering structures by projecting artificial textures onto them. The possibility of the creation of models of the subject increases the accuracy and controllability of the result. With increasing contrast and complexity of the texture being applied, digital photogrammetry algorithms better determine the correspondence between a pair of photographs.