Novel Technique for IRFPA Blind Pixels Detection

Abstract

Infrared focal plane array (IRFPA) is a bi-dimensional infrared detector which is a matrix of micro detectors referred as pixels. Physical pixel failure (blind pixel) is a serious issue in case of utilizing IRFPA as a measuring tool or a surveillance system to monitors hot objects. Scene-based non-uniformity correction (SBNUC) technique for uncooled IRFPA depends on the correction algorithm strength to estimate the correction parameters and to eliminate the need for using the reference-based non-uniformity correction (RBNUC) technique which requires doing a lab calibration instead of self-calibration. SBUNC’s have disadvantages of ghosting artifacts which be complex when a new blind pixel appears. This research aims to enhance SBNUC techniques by changing the detection technique blind pixels and apply it in an earlier stage of system operation and to eliminate the need for adding a complex layer of the algorithm over original SBUNC codes for blind pixels detection, only. This research focuses on studying the electrical aspects of IRFPA detector to examine the effect of skimming reference voltage on the global image contrast. This research found out that exciting Vsk by voltage values close to its boundaries could separate blind pixels from the image background which become very exposed and simple to detect and correct. This research proposes a combination of electrical circuit modification, a routine of Vsk voltage excitation along with image acquisition. This research applies Otsu’s method as a global thresholding technique to detect blind pixels. Researchers found out that this proposed technique could detect all blind pixels successfully without the need for RBNUC which avoid SBNUC's techniques of lab calibration process to detect the blind pixels.