RECOGNITION OF ROAD OBJECTS AS A MEANS OF SOLVING THE PROBLEM OF INATTENTION ON THE ROAD

Abstract

The subject of the article is the problem of inattention on the road and its root cause. The topic of this article is methods for recognizing objects on the road in real-time. The tasks of this article are to investigate the statistics of the causes of road accidents, to analyze the capabilities of modern methods of object recognition and computer vision, and to analyze the possibility of integrating these methods and models into the smartphone application. The end goal of research and analysis is to try to develop software that can help visually impaired and inattentive drivers, as well as reduce the potential number of road accidents. In this article used next methods: contour analysis methods with Fourier transforms and convolutional neural network design and training methods. The results of the analysis of road accidents causes showed that the main problem on the road - is the inattention of drivers. The analysis of models, methods and possibilities of their integration showed that modern methods of object recognition are capable of solving the task, but there are many factors that affect the accuracy of these methods. Recognition accuracy is also highly dependent on the smartphone and its processing power, camera, etc. Conclusion. Achieving the goal is possible, but modern algorithms and smartphones can solve it only partially. Also, we do not really know if it potential developed app will help driver and would not confuse them more on the road.