Performance evaluation of image enhancement techniques on a digital image-intensifier Judith Dijk ∗ , Piet Bijl § , Henri Bouma ∗ ∗ TNO Defence, Security and Safety, P.O. Box 96864, 2509 JG, The Hague, The Netherlands email: Judith.Dijk@tno.nl § TNO Defence, Security and Safety, P.O. Box 23, 3769 ZG, Soesterberg, The Netherlands ABSTRACT Recently new techniques for night-vision cameras are developed. Digital image-intensifiers are becoming avail- able on the market. Also, so-called EMCCD cameras are developed, which may even be able to record color information about the scene. However, in low-light situations all night-vision imagery becomes noisy. In this paper we evaluate the performance of image enhancement techniques for one type of noisy night imagery, that is a digital image-intensifier. The image enhancement techniques tested are noise reduction, super-resolution reconstruction and local adaptive contrast enhancement. The results show that image enhancement tech- niques improve the usage of image-intensifiers in low-light conditions. The largest improvement is found for super-resolution reconstruction applied at the smallest objects. This indicates that part of the improvement is obtained by resolution enhancement. Applying LACE does not change the performance, indicating that in this setup LACE performs equal to the automatic gain control of the image-intensifier. Keywords: Image enhancement, super-resolution, contrast enhancement, image-intensifiers, TOD 1. INTRODUCTION For all military operations, situational awareness is of great importance. This situational awareness can be obtained by using cameras. The current trend is that more and more operations are shifted from daytime to night. This increases the need for night-time imagery. In night time operations, image-intensified cameras are used next to infrared cameras. The advantage of image-intensified cameras with respect to infrared cameras is that they allow for visual identification. Low light levels are normally not sufficient to see details of the scene. The image-intensifier uses light from the stars or the moon to obtain an image under low-light conditions. This image enables the soldier to interpret his environment. Traditionally, image-intensifiers are carried by the individual soldier. Traditionally, image-intensifiers are analogue devices. The intensified image is only shown to the soldier wearing the device. Currently digital image-intensifiers are being developed. Using these cameras a number of capabilities are possible that analogue image-intensifiers lack. Among the new possibilities are 1) showing the images to other users, 2) enhancement of the intensified imagery 3) fusion of the image-intensifier image with other imagery (for instance infra-red) 4) adding context information to the imagery and 5) fusion of the information with data from other sensors, e.g. in a net-centric environment. This makes the quality of the digital imagery a topic of interest. One of the differences between a digital and analogue system is its noise behavior. The noise in an image-intensifier can be modeled as Poisson noise, with its characteristic peaks. In an analogue system this effect will be visible in the individual fibers, whereas for a digital system the peak noise will be distributed over a pixel, which has a larger footprint than the fibers. In digital images, however, the noise effects can be reduced by the use of image enhancement. In this paper, we evaluate the performance of image enhancement techniques for image-intensified imagery. The image enhancement techniques that are applied are described in section 2. The evaluation is done using the TOD method. Some details about this method are given in section 3. The experimental setup including details about the recordings is given in section 4. The results of the experiments are presented in section 5. Conclusions and directions for further research are presented in section 6. Infrared Imaging Systems: Design, Analysis, Modeling, and Testing XX, edited by Gerald C. Holst Proc. of SPIE Vol. 7300, 73000F · © 2009 SPIE · CCC code: 0277-786X/09/$18 · doi: 10.1117/12.820002 Proc. of SPIE Vol. 7300 73000F-1