Improving Unet segmentation performance using an ensemble model in images containing railway lines

Abstract

This study aims to make sense of the autonomous system and the railway environment for railway vehicles. For this purpose, by determining the railway line, information about the general condition of the line can be obtained along the way. In addition, objects such as pedestrian crossings, people, cars, and traffic signs on the line will be extracted. The rails and the rail environment in the images will be segmented with a semantic segmentation network. In order to ensure the safety of rail transport, computer vision, and deep learning-based methods are increasingly used to inspect railway tracks and surrounding objects. In particular, the extraction of objects around the railway line has become an important task. The dataset contains images of the railway line and its surroundings, which were obtained in changing environmental conditions, at different times of the day, and under poor lighting conditions. In this study, a new method is proposed for the extraction of objects in and around the railway line. The proposed approach first applied Unet-based segmentation methods on the dataset. Then, a method that improves Unet performance based on the ensemble model is proposed. ResNet34, MobileNetV2, and VGG16 backbones were used to improve segmentation performance. The proposed model is based on the ensemble decision-making process, significantly contributing to the semantic segmentation task. Experimental results of the developed model show that it gives 85% accuracy rate and 54% average IoU results.