A Theoretical Framework Towards Building a Lightweight Model for Pothole Detection using Knowledge Distillation Approach

Abstract

Despite recent advances in deep learning, the rise of edge devices, and the exponential growth of Internet of Things (IoT) connected devices undermine the performance of deep learning models. It is clear that the future of computing is moving to edge devices. Autonomous vehicles and self-driving cars have leveraged the power of computer vision, especially object detection to navigate through traffic safely. Nevertheless, to be able to drive on all types of roads, these new vehicles have to be equipped with a road anomaly detection system, which strikes the need for small deep learning models of detecting road anomalies that can be deployed on these vehicles for safer driving experience. However, the current deep learning models are not practical on embedded devices due to the heavy resource requirements of the models, as such cannot be deployed on embedded devices. This paper proposes a theoretical approach to building a lightweight model from a cumbersome pothole detection model that is suitable on edge devices using knowledge distillation. It presents the theoretical approach of knowledge distillation, why it is a better technique of model compression compared to the rest. It shows that a cumbersome model can be made lightweight without sacrificing accuracy and with a reduced time complexity and faster training time.

Image-Based Pothole Detection Using Multi-Scale Feature Network and Risk Assessment

“…However, real-time object detection cannot be guaranteed in embedded systems with low-performance GPUs and CPUs, even if the computation speed of YOLO is fast [11]. Therefore, the YOLO-tiny series algorithm, advantageous for real-time detection in embedded systems, has been widely applied to pothole detection [12]. However, YOLO-tiny series algorithms have lower accuracy than general YOLO algorithms because the network size is reduced for computation speed.…”

Fala: An intelligent mobile application for safe driving

“…Pothole is one of the challenging effects of safe driving as it portrays a lot of dangers to road users. A pothole is a natural cave or a hollow on the road surface formed due to erosion or aging of asphalt [1]. Potholes pose a lot of dangers for road transport users in many developing countries, especially in Nigeria.…”

Computer vision for asphalt cracks detction using YOLOv5

“…Therefore in this research this two terms will be interwoven. A pothole is a natural cave or a hollow on the road surface formed as a result of erosion or aging of asphalt [1]. Potholes pose a lot of dangers for road transport users in many developing countries, especially in Nigeria.…”