研究テーマResearch theme

In this study, we mainly carry out deep learning based road marking detection method and classification method for damage degree calculation.

研究内容の概要Overview

This study mainly focuses on the problem of road sign damage detection and damage degree classification, and carries out research on the uncertainty road sign detection method based on the combination of deep learning and fuzzy mathematics. We explore the network structure suitable for road signage high-precision detection network, and construct a damage classification model based on this network structure combined with fuzzy mathematical methods, in order to improve the efficiency of road signage restoration.
The main research contents are as follows:
(1) Research on accurate detection method of road marking based on deep learning method.
(2) Research on the classification method of road marking damage degree based on fuzzy mathematics method.
(3) Construction of road marking damage detection and classification method based on the fusion of lightweight network and fuzzy mathematics.

研究成果をどのように社会に役立てるか
（還元の構想）Giving back to society

The research method of uncertainty road marking detection based on the combination of deep learning and fuzzy mathematics is attempted to apply the deep learning detection method and fuzzy mathematics on the detection of the damage degree of the road marking, which can improve the efficiency of the road marking repair, save a great deal of manpower and physical costs, and reduce the accidents due to the unclear road marking. Also, the method is applicable in other tasks of uncertainty and model label recognition. This method can detect the wear and tear of road markings efficiently, quickly and easily. It can greatly reduce the labour cost and equipment cost.

研究者一覧へ戻る