A Research On Improving The Dynamic Range Of An Area-array CCD Imaging System

Natural scenes can emit a wide range of luminance:the starlight has luminance of about0.001cd/m2at night while the shining sunlight has luminance of about1,000,000cd/m2at noon. The dynamic range can be up to nine orders of magnitude. Due to the automatic adjustment of human visual system, people can discriminate intensities spanning a dynamic range of about110dB. Unfortunately, ordinary imaging sensors can only capture a dynamic range of about60dB because of the limits of charge coupled devices and analog-to-digital converters. Therefore, images captured by these sensors cannot reflect the real scenes. Research about expansion of image dynamic range has a technological and commercial development vista. Apart from that, it also has a great significance in spaceborne and airborne industry, military reconnaissance and etc.Existing expanding method can be divided into two categories:hardware expansion and software expansion. The hardware expansion method requires increasing charge capacity for each pixel, enhancing the photoelectric sensitivity during photosensitive period, and addressing the resulted enlarged noise signals. This technology is complicated and too expensive to be widely used. On the other hand, although the software method is much more cheap, fast and easy to achieve, it is post processing and therefore cannot output the image with high dynamic range in real time.Aimed at solving the problem mentioned above, a CCD imaging system which can output real-time high dynamic range images is designed in this paper. CCD chip ICX415AL from SONY Co. is used as an imaging sensor and the required drive signals are generated by FPGA. By controlling electronic shutter signal-SUB, two adjacent images have different exposures. Based on the feature of human visual system, each pair of adjacent images is combined using a synthesis algorithm carried out by FPGA in order to obtain real-time high dynamic range images.This paper consists of i) an imaging system’s circuit design using CCD drive signals and sampling clock signals both provided by EP1C12F256; ii) a multiple exposure method which makes adjacent images with different exposure time and preserving different parts of the luminance region so that we can expand the dynamic range by combining the two images together; and iii) a synthesis algorithm based on human visual system model which makes the processed images more detailed and more approximate to human visual preferences.