5 May 2017 Block match denoising for the Integrated Digital Vision System (IDVS)
Author Affiliations +
There is a need to develop fast vision systems capable of supporting real time operations that require split-second decision making. To perform at high speed, these vision systems are subject to stringent latency requirements thus hindering their light sensing elements to collect a meaningful number of photons with an acceptable Signal to Noise Ratio (SNR). As a result, high amplifier gains end up amplifying large amounts of noise along with image content. Rockwell Collins developed an all-digital vision system, dubbed Integrated Digital Vision System (IDVS) with very low latency capable of operating real time in conditions ranging from complete darkness to daylight. This paper presents an algorithmic approach to denoise IDVS frames based on state of the art image denoising algorithms including Block Matching 3D (BM3D) and Non Local Mean (NLM) algorithms that are modified to meet IDVS hardware restrictions.
Conference Presentation
© (2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Mokhtar M. Sadok, Mokhtar M. Sadok, John S. Alexander, John S. Alexander, Andrew J. LeVake, Andrew J. LeVake, } "Block match denoising for the Integrated Digital Vision System (IDVS)", Proc. SPIE 10197, Degraded Environments: Sensing, Processing, and Display 2017, 101970G (5 May 2017); doi: 10.1117/12.2264241; https://doi.org/10.1117/12.2264241


Back to Top