Title
Automated real-time correction of exposure times in a multispectral camera using an ambient light sensor
Abstract
Maintaining appropriate exposure times is challenging in applications that involve continuous capturing and processing of images in an environment of dynamically changing light. This problem is typically solved by post processing the captured images and adapting the exposure time for the next image based on some brightness metric derived from a previous image. However, this results in a delay between the light changing and the exposure time being properly updated, resulting in under- or overexposed images. In this article we propose using a light sensor to actively measure the changing light and use this to calculate the appropriate exposure time prior to images being captured. We show that by using this method we can capture images that are appropriately exposed for maximum contrast even when the ambient light changes dramatically.
Year
DOI
Venue
2014
10.1145/2683405.2683458
IVCNZ
Keywords
Field
DocType
design,digitization and image capture,reliability,local binary pattern,nakagami distribution,support vector machine
Computer vision,Pattern recognition,Computer science,Multispectral image,Support vector machine,Local binary patterns,Image based,Photodetector,Nakagami distribution,Artificial intelligence,Brightness
Conference
Citations 
PageRank 
References 
0
0.34
1
Authors
3
Name
Order
Citations
PageRank
Jaco Fourie100.34
Kenji Irie200.34
Jessica Roberts322.85