Design Of DAC And Algorithm Research Of Image Denoising For The Improvement Of Visual Effects Of Images

In this information society, imaging has become an important means of obtaining information, expressing information and delivering information. It is the basis for human to perceive the world. However, images often consist of much useless information and can not be used directly. Fortunately, the development of image denoising technology helps people solve the problem, it has been widely applied in many observation fields such as microwave measurement, biological observation and so on.Digital-to-Analog Converter (DAC) is an important part of image sensor. Denoising performance of image sensor has significant influence on overall observation system performance. There are two ways to improve anti-noise performance of observation system and improve visual effects of acquired observation images:(1) Internal circuit design improvement in image sensor, for example DAC module, should be improved to have higher accuracy; (2) Software processing is applied for the images acquired from observation system. Corresponding to the two directions, the following research work has been done:(1) In order to improve the precision of DAC further, fibonacci sequence decoding is utilized in the design of current steering DAC. A segmented scheme whose upper bits use unary sequence and lower bits use binary sequence is usually adopted to balance area and accuracy. Different from the traditional method, in this work,6-bit Fibonacci sequence is used to replace the lower 6-bit binary sequence in designed 12-bit current steering DAC.(2) For the improvement of traditional median denoising algorithm, a switch median denoising algorithm based on linear dependency thresholds has been proposed. The algorithm designs a self-adjusting strategy to adjust algorithm parameter and deals with boundary problem. It has been proved that it makes better denoising effects.(3) Combined with the concept of information entropy in image denoising processing, an entropy dropping denoising algorithm based on the conception of information entropy has been proposed. The algorithm firstly constructs image feature space, then entropy dropping is measured by the function of Parzen Window density estimation, finally the algorithm is stopped by termination strategy. It has been proved that the algorithm makes better denoising effects.