Kirill D. Bessonov, Kirill K. Tomchuk

 Saint Petersburg State University of Aerospace Instrumentation (SUAI) (St. Petersburg, Russia) 

 bessonovkirill1997@ya.ru,  tomchuk@guap.ru

Examined video camera which is explored in the present article includes following main components: CMOS (complementary metaloxide-semiconductor)-matrix ICX415AL and analog video amplifier (AMP) AD9824KCP which change its gain with different input binary code. Exposure (A) control is implemented programmatically in the main FPGA (Field Programmable Gate Array) and has range within 621. Max gain of AMP (G) is 36dB which corresponds to 1023 number. Changing these two parameters one can change mean brightness (I) of received image. By changing A and G independently and registering I 3-D surface can be plotted. The idea of first method was that the form of I variation will not change over different outer illumination level (I0). After plotting of three surfaces with different I0 author saw significant surface deformation. With such deformation, no surface approximation can be made. But this experiment showed the common form of f(G,A) dependence. This information helped simulate the second method more accurately. The second method uses linear approximation for finding such G and A values that meet optimal I (Iopt). The main distinction  between first two methods is that G and A values in the second method are changing sequentially. This approach helps dispose of three-dimensionality of f(G,A) function. Second algorithm vary A first and when it reached its upper limit it vary G because AMP  amplifies both the main video signal and noise. The simulation showed that Iopt could be met with no more than six steps (of six frames). If by some reasons linear approximation method does not work, then the whole approximation instrument could be replaced with incrementation/decrementation of G and A values. This approach is easier for software implementation, but has larger time consumption because every incrementation/decrementation procedure will take one frame for execution. However, in the low data speed systems, such as video surveillance, this approach could be used.

Bessonov К.D., Tomchuk К.К. Automatic regulation of camera parameters based on approximated image intensity function. Achievements of modern radioelectronics. 2020. V. 74. № 11. P. 63–67. DOI: 10.18127/j20700784-202011-11. [in Russian]

1. Kao W.C., Hsu C.C., Kao C.C., Chen S.H. Adaptive Exposure Control and Real-Time Image Fusion for Surveillance Systems. IEEE International Symposium on Circuits and Systems. 2007. № 1. P. 935–938.
2. Battiato S., Castorina A., Mancuso M. High dynamic range imaging for digital still camera: an overview. Journal of Electronic Imaging. 2003. № 3. P. 459–469. 
3. Mantiuk R., Krawczyk G., Myszkowski K., Seidel H. Perception-motivated High Dynamic Range Video Encoding. ACM Trans. Graphics. 2004. № 3. P. 733–741. 
4. Habili N., Lim C.C., Moini A. Segmentation of the Face and Hands in Sign Language Video Sequences Using Color and Motion Cues. IEEE Trans. Circuits and Systems for Video Technology. 2004. № 8. P. 1086–1097. 
5. Challapali K., Brodsky T., Lin Y.T., Yan Y. Real-Time Object Segmentation and Coding for Selective-Quality Video Communications. IEEE Transactions on Circuits and Systems for Video Technology. 2004. № 6. P. 813–824.
6. Kao W.C., Chen S.H., Sun T.H., Chiang T.Y. An Integrated Software Architecture for Real-Time Video and Audio Recording Systems. IEEE Trans. Consumer Electronics. 2005. № 3. P. 879–884.
7. Stark J.A. Adaptive Image Contrast Enhancement Using Generalizations of Histogram Equalization. IEEE Transactions on image processing. 2000. № 5. P. 889–896.
8. Zhihui G., Yuanbao G., Hongtao Y. Auto-Exposure Algorithm Based on Luminance Histogram and Region segmentation. Applied Mechanics and Materials. 2014. № 1. P. 2278–2282.
9. Matthews S., Al-Janabi K. Automatic Exposure Control (AEC) for EIS. Mullard space science laboratory. 2003. №MSSL/SLBEIS/SP027.02. P. 1–7.
10. Shian J., Fuh C.S. Automatic Exposure with Center/Focus Metering. Proceedings of Conference on Artificial Intelligence and Applications. 2005. № 1. P. 56.