傅楸善臺灣大學:資訊工程學研究所邱立誠Chiu, Li-ChengLi-ChengChiu2010-06-092018-07-052010-06-092018-07-052009U0001-3006200918395500http://ntur.lib.ntu.edu.tw//handle/246246/185359The technology of digital still camera (DSC) is a popular field recently. With the growth of demand, low cost and high quality are the main trends in the consumer DSC market. To avoid increasing hardware cost, all kinds of DSC algorithms are explored and developed by the specialists and experts in this field. In our research, we focus on three main technologies of DSC, color restoration, auto focus, and de-noise. o deeply explore and understand the architecture of DSC, we implement a software simulator including full functions of DSC. This simulator can accept raw data from DSC and output final processed Joint Photographic Experts Group (JPEG). It is not only convenient for image quality adjustment but also easy for algorithm comparison.We design color restoration, auto focus, and de-noise algorithms. Our color restoration uses two-step procedure to detect environment color temperature and correct the color by such information. Proposed auto focus adopts a prediction model which is derived from optical and signal processing to shorten overall search time. Our de-noise filter is inspired by the well-know bilateral filter. It can perform similar noise removal but achieve better edge preservation.1 Introduction ………………………………………………………….. 1.1 Overview ……………………………………………………………………….. 2.1.1 Image Signal Processor Simulator ……………………………………... 2.1.2 Color Restoration ………………………………………………………. 2.1.3 Auto Focus ……………………………………………………………... 3.1.4 De-noising ……………………………………………………………… 3.2 Thesis structure ………………………………………………………………… 4 ISP Simulator ………………………………………………………… 5.1 Simulator Structure …………………………………………………………….. 6.1.1 Raw data import ………………………………………………………... 7.1.2 Optical black …………………………………………………………… 9.1.3 White balance …………………………………………………………. 10.1.4 Color interpolation ……………………………………………………. 12.1.5 Color correction ………………………………………………………. 13.1.6 Gamma correction …………………………………………………….. 14.1.7 Edge enhancement ……………………………………………………. 15.1.8 Processed image data export ………………………………………….. 16.1.9 Control of image pipeline …………………………………………….. 16.2 Simulation Results ……………………………………………………………. 17 Color Restoration …………………………………………………… 22.1 Problem Formulation and Previous Related Works …………………………... 23.2 Our Proposed Method ………………………………………………………… 26.2.1 White balance algorithm ……………………………………………… 27.2.2 Color correction algorithm ……………………………………………. 33.3 Experimental Results …………………………………………………………. 34.3.1 White balance evaluation ……………………………………………... 34.3.2 Color restoration simulation ……………….…………………………. 39.3.3 Practical implementation of color restoration ………………………… 42 Auto Focus …………………………………………………………… 44.1 Focus Value Curve Formulation and Time Consumption Evaluation ………… 47.1.1 Optical lens system …………………………………………………… 47.1.2 Focus value curve …………………………………………………….. 49.1.3 Noise effect and focus value ratio ……………………………………. 53.1.4 Lens step versus motor rotation rate ………………………………….. 55.2 Our Proposed Predictive Auto Focus Algorithm ……………………………... 56.3 Experimental Results …………………………………………………………. 59 De-noising …………………………………………………………… 65.1 Review of Bilateral Filter ……………………………………………………... 67.2 Proposed De-noising Filter …………………………………………………… 69.3 Experimental Results …………………………………………………………. 73 Conclusions and Future Works ……………………………………. 81.1 Image Signal Processor Simulator ……………………………………………. 81.2 Color Restoration ……………………………………………………………... 82.3 Auto Focus ……………………………………………………………………. 83.4 De-noising …………………………………………………………………….. 84 Bibliography ………………………………………………………… 85 Publication Lists ………………………………………………….… 9318074873 bytesapplication/pdfen-US數位相機色彩還原自動對焦雜訊去除軟體模擬器digital still cameracolor restorationauto focusde-noisesoftware simulatorthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/185359/1/ntu-98-D93922010-1.pdf