Abstract:

The evolving of depth camera with deep learning technology have been setting the pavement for various applications in today’s digital world where realtime depth detection and distance estimation is one of them. The utilization of depth data allows us to estimate the depth and object’s distance to overcome the tailgating detection, fatal accidents, object collision etc. in autonomous vehicles. This paper focuses on real-time depth detection using both geometric computation and rule-based algorithms. The proposed method aims to overcome the problems of existing depth detection techniques such as scaling issues, object overlap etc. by providing good and effective depth estimation from realtime video stream images. The objective of this work is to evaluate real-time depth detection to identify the obstacles and plan safe paths to avoid collision in autonomous vehicles. Finally, the proposed hybrid approach not only provides a limited degree of accuracy but also a computational speed which is crucialfor realtime applications. Experimental results illustrate that the proposed method surpasses traditional techniques in terms of both performance and reliability, demonstrating its potential to be an alternative solution in the field of real-time depth detection.

Keywords: Depth detection, geometric computation, rule-based algorithm, autonomous vehicles

Conference Name: 10^th international Conference on Business Analytics and Intelligence (2023- ICBAI)