• Proceedings of the KSRS Conference
• /
• v.2
• /
• pp.788-790
• /
• 2006

The algorithm of calculating NUC table, based on Image data collection, is based on two basic assumptions. These basic assumptions are as follow: one is the NUC is of a linear nature. The other is all pixel see the same statistical distribution for large number of lines. We generated NUC tables for a radiometric calibration & validation of KOMPSAT-2 using a dark cal. Data.

• Aerospace Engineering and Technology
• /
• v.11 no.1
• /
• pp.98-102
• /
• 2012

These days, Uncooled IR Systems are more popular in the area of defense and aerospace than before. Uncooled IR Systems are widely used as core technology for making unmanned systems and detecting enemy objects during the day and night in the distance. Recently, researches on TEC-less IRFPA have been increased to minimize the power consumption and to make a smaller system than before. For this, it needs to find adequate NUC(Non-Uniformity Correction) coefficients as FPA(Focal Plane Array) temperature changes. In this paper, we propose a new NUC coefficient estimating technique, DCCE-LI(Dynamic Calibration Coefficients Estimation with Linear Interpolation), for TEC-less IRFPA. It is based on a linear interpolation method and it can estimate NUC coefficients in real-time. So, by testing and evaluating it with some IR images, we conclude that the quality of IR images using proposed method is better than applying static coefficients.

• Korean Journal of Remote Sensing
• /
• v.27 no.2
• /
• pp.89-98
• /
• 2011

Four kinds of gain and offset correction coefficients that are used to correct the nonuniformity between pixels are discussed. And their correction performance has been compared by performing image correction. using the correction coefficients calculated, on the real image data obtained from a newly fabricated camera electronics system. The performance of the correction coefficients depends in general on the number of the light input levels used to obtain the reference image. The result shows that, as expected obviously, when only two light input levles are used to obtain the reference image, even though its correction coefficients are relatively easily calculated, the correction performance is relatively poor. And with the number of light inputs increased to a value of larger than two, the correction performance is improved. It is noted, however, no Significant performance difference is found between the different correction coefficients employed.