Research Of High Dynamic Range Imaging Based On FPGA

Dynamic range refers to the ratio of the maximum value and minimal value of a physical quantity. In the field of digital imaging technology, when the range of illumination of the scene to be captured extends too large, traditional camera may fail to record the scene correctly due to its limited capability of sensing range. The bright area may be overexposure and the dark area may be underexposure, both situation causes loss of detail information of the scene. The High Dynamic Range(HDR) imaging provides a good solution to this problem. This technology allows us to precisely record the all the information in different areas with different illumination, and produces more natural and higher quality images that is consistent with the human perception of high dynamic range scene, which can also be used in the field of image processing and computer vision. HDR imaging has been widely used in many applications such as video surveillance, satellite remote sensing, medical imaging and so on.There are two major ways to obtain HDR image. The first way is to use specially designed image sensors, but this may increase the complexity and cost of hardware, therefore could not be widely used. The second way is to combine multi-exposure images captured by traditional camera. There is no requirement for specialized hardware and could be easily applied in many fields.This paper introduces and studies the high HDR technology, the main work of this paper is shown below:(1) Introducing HDR imaging technology, mainly focuses on the multi-exposure method. Illustrating each steps of the technology, including camera response function recovery, HDR image reconstruction, Tone mapping technology, image fusion technology, and HDR image storage.(2) Studying the exposure control algorithm for HDR imaging. The first step of HDR imaging is to obtain pictures with different exposure. First introducing the common exposure bracketing and the content-based multi-exposure algorithm. To overcome their drawbacks, an exposure control algorithm based on luminance area division is proposed in this paper. It divides the image into different luminance area by using OTSU thresh-holding, and covers detail information in different areas.(3) Designing and building the image acquisition system based on FPGA platform for support of HDR imaging. Through this platform the exposure control of the image sensor can be achieved in real time. Cooperating with the software running on PC, HDR image could be obtain and shown.(4) Experiments for generating HDR images is executing on the platform. Experimental results show the validity and advantage of the proposed exposure algorithm by comparing to the common exposure bracketing. Also experiments for exposure of ROI( Region of Interest) is done. Results shows that using the HDR tech to capture human face images under non uniform illumination can effectively eliminate the influence of non-ideal light condition.