Development of Road Detection Algorithm on Mobile Cart Platform

Junkwang Kim, Young-Duk Kim


Recently, research related to mobile deep learning research and autonomous driving has been actively conducted in various fields. In particular, the lane detection algorithm secured high performance and real-time performance by using deep learning research. However, it is very difficult to apply the lane detection algorithm to an environment with various characteristics that do not have a uniform shape, such as a golf cart road. In addition, since it is difficult to utilize a high-performance chip in the small golf cart platform, it is important to develop a fast and lightweight mobile deep learning research. In this paper, we propose a cart road detection algorithm for a small mobile platform. As a result, based on the Resnet-50 network, we achieved IoU 88.4%, top1 accuracy 82.1%, top3 accuracy 97.6%, and 5.3fps.


deep learning; mobile application; robot software; embedded software; detection

Full Text:



  • There are currently no refbacks.