Improving tracking accuracy using illumination neutralization and high dynamic range imaging

Abstract

Object recognition and tracking is a common task in video processing with multiple applications including surveillance, security, industrial inspection, medicine, and more. Recognition and tracking accuracy can drop significantly when the scene dynamic range exceeds that of common camera sensors, which may lead to decreased tracking accuracy. The problem is further exacerbated when the scene illumination changes abruptly, such as when a light is turned off, or when clouds obscure the sun. This chapter investigates high dynamic range (HDR) imaging as a possible solution to these problems. HDR imaging extends the accuracy of the captured image, which naturally improves the performance of existing algorithms, while further gains are possible when algorithms are designed to specifically take advantage of HDR information. This chapter discusses an algorithm that improves the accuracy of object tracking algorithms by employing a custom HDR capture methodology and an image neutralization algorithm designed to take advantage of HDR information. © 2017 Elsevier Ltd. All rights reserved.