Paper Title :Robust and Real-Time Road Detection on Indian Roads using Deep Learning

Author :Kojagiri Kakade, Suraj Sawant, Prasenjeet Damodar Patil

Article Citation :Kojagiri Kakade ,Suraj Sawant ,Prasenjeet Damodar Patil , (2022 ) " Robust and Real-Time Road Detection on Indian Roads using Deep Learning " , International Journal of Advance Computational Engineering and Networking (IJACEN) , pp. 7-12, Volume-10,Issue-9

Abstract : Abstract - In recent times, tremendous work has been done in the field of road detection. A panoptic perception system can help in road detection and assist the navigation of vehicles. With the help of real-time visual information, a training-based road detection approach is followed in this paper. The custom dataset is constructed from the frames of captured road videos collected using different mobile platforms. The road videos are collected from different geographical locations in India and under different environmental conditions. This custom dataset is annotated for different perception tasks like object detection, lane detection, and road segmentation using the open source Scalabel annotation tool. Here, all the images of the dataset were trained and tested using the multi-task algorithm. Multi-task based learning approach is used to perform different perception tasks simultaneously. With the help of experimental results, we can evaluate the effectiveness of different multi-task algorithms on the custom dataset. Keywords - Deep learning, dataset, YOLO, object detection, lane detection, area segmentation, end-to-end network multi-task.

Type : Research paper

Published : Volume-10,Issue-9

	|		PDF	|	Viewed - 43	|	Published on 2022-12-21