Locating Intersections for Autonomous Vehicles: A Bayesian Network Approach

Abstract

A novel idea is presented to locate intersections in a video sequence captured from a moving vehicle. More specifically, we propose a Bayesian network approach to combine evidence extracted from a video sequence and evidence from a database, maximizing evidence from various sensors in a systematic manner and locating intersections robustly.

Road intersection detection and classification using hierarchical SVM classifier

“…Many measures are incorporated for good evaluation of the proposed solution. The obtained results are compared to those of Choi et al (2007).…”

“…The Bayesian Network (BN) proposed by Choi et al [11] combines the evidence from a database and a computer vision engine for road intersection detection. This approach shows a higher recognition rate compared to the obtained rate when using only visual information.…”

“…Its main drawback is the need of a priori knowledge about the road network. In its actual form, [11,33] the BN allows the detection of road intersection without searching for their types.…”

“…The proposed solution is compared to the approach presented in [11]. For this purpose, we address firstly the problem of road boundary detection and road form extraction.…”

“…Currently, the number of papers addressing this problem is very limited. [11][12][13] In [11], the evidence of the detected road intersection from visual data has been combined with evidence from database using a Bayesian Network (BN) approach. The road intersection recognition rate achieved using this approach is of 84%.…”