A Heatmap-Based Approach for Analyzing Traffic Sign Recognition and Lane Detection Algorithms

Abstract

A heat map (or heatmap) is an efficient and convenient way to condense the information into a single snapshot. It can also be used in the analysis of high-dimensional data structures that are representative of large data sets generated by vehicle sensory systems. In this paper, we show how to generate such heat maps from vision data recorded by the front camera in the vehicle. We present two algorithms that can be applied to build heat maps that represent the detection status of traffic signs and road lanes. The resulting graphs can then be used to evaluate in a qualitative and quantitative way the Traffic Sign Recognition and Lane Detection functionalities that are part of Advanced Driver Assistance Systems. The effectiveness of the presented approach is illustrated in several use cases of applications.