Research On Key Technology Of Digital Camera Imaging System

With the development of digital technology, digital imaging devices have beenwidely applied to our life, which causes the tendency to study on this field. From theperspective of theory and practicability, the researches on digital imaging field arevaluable and promising. Imaging quality is the vital issue of the researches on digitalimaging devices. As the process of imaging is influenced by plenty of factors, theoutput images are usually hardly consistent with the real scene seen by human eyes.Therefore, how to improve the imaging quality has been becoming the core of thisresearch field.Exposure, focusing and white balance of digital camera are the important parts ofimaging process. The imaging quality is affected directly by these three factors.Nowadays, researchers at home and abroad have paid many years on these technicalstudies which have achieved success at some aspects. However, as the operatingenvironment is becoming diversified, many technical problems still need to befocused on and improved. Because of that, how to ameliorate the performance ofdigital camera on exposure, focusing and white balance has both theoretical value andpractical perspective. In this dissertation, the specific work is managed as follows:From the physiological structure of human eyes, this dissertation firstly analyzesthe photoreceptive principles of human vision system and elucidates some visualfeatures of human eyes. Secondly, through in-depth researches on imaging system ofdigital camera, dig out a series of important factors associated with imaging qualityand analyzes the effects of exposure, focus and white balance of digital camera. Also,we seek for some effective methods to advance the performance of digital colorcamera by comparing the human vision system with camera imaging system, whichwill lay the foundation for subsequent research.Based on the photoreceptive features of human vision system, this dissertation presents an automatic exposure method for color digital camera to solve theunderexposure or overexposure problems of unmanned camera under some specialconditions, especially for high dynamic range scenes or small targets scenes. Thismethod contains two parts: first of all, construct an exposure evaluation functionusing the image information about sharpness, saturation and luminance, which canimprove the accuracy and robustness of exposure evaluation, compared with theevaluation method only using scene luminance; what’s more, construct an adjustmentcurve between the exposure evaluation values and exposure gain, which reduces theframe number needed for exposure adjustment and increases the efficiency ofautomatic exposure.Meanwhile we make researches on the principles of automatic focus of digitalcamera and ameliorate a focus-evaluation algorithm. Considering the color correlationamong different pixels, a chromatic aberration factor is introduced into evaluationfunction based on Brenner sharpness evaluation. This method solves the matter ofevaluation failure caused by transformation among different color spaces. Thenon-linear coefficient is used for improving the stability of evaluation algorithm. Atthe same time, for the small objects in moving scenes, propose an automatic focusmethod by the use of optical flow estimation of moving targets. This method canidentify interested objects in a multi-target scene accurately and adjust focusparameters based on these targets under some complex-background condition, whichmaintains the stability and accuracy of focus.According to the Retinex Theory, an advanced automatic white balance methodis proposed. This method takes advantage of image color information to divide thewhole picture into the simple-color regions which represent the image color style, andthe rich-color regions which express detail information of a scene. By this way, theinfluence of simple-color region on image color temperature evaluation can bedecreased and the robustness and applicability of the algorithm can be increased. Atthe same time, considering the limitation of subject-evaluation on color image, this dissertation proposes an objective method for evaluating white balance. In CIEchromaticity coordinate, the color property of each pixel and its relative locationinformation in the chromaticity coordinate are utilized to construct an imagecolor-barycenter model. As an evaluation method of post-imaging, this method whoseevaluation consequence is consistent with human vision system’s is able to evaluatecolor balance for different images accurately and objectively. It could provide areference standard for post-process.