Image Enhancement and Noise Model Research Based on Scatter
|School||South China University of Technology|
|Course||Circuits and Systems|
|Keywords||Systematic errors Scattered noise Dark primary colors priori Energy adjustment Image Enhancement|
In order to provide better image input, image preprocessing is an essential step in image technology. The systematic error is one of the main reasons of image quality, in this article from the point of view of systematic error study the imaging system error and provide an effective error cancellation algorithm in order to achieve the purpose of the recovery and enhanced image. This paper first introduces the commonly used method of image pre-processing, and pointed out that these methods to eliminate the lack of systematic errors; overview of common systematic errors in several imaging systems, such as atmospheric scattering ray scattering, and quantum noise lens MTF, image quality evaluation criteria given from the four aspects of the noise, contrast, detail and sensitivity; discussed in detail various systematic errors, error elimination. Image fog exists, through the study of atmospheric scattering, systematic introduction to the atmospheric scattering model, the introduction of a priori knowledge of the dark primary colors, elaborate the demisters algorithm based on the prior knowledge of a single image. In order to improve the efficiency of the algorithm proposed in this paper based on the dark a priori primary colors of the rapid demisting algorithm, the experimental results show that the proposed algorithm to ensure the the defogging quality while greatly improves the efficiency of the algorithm can achieve real-time processing requirements. Ray image boundaries fuzzy-ray scattering studies, quantitative analysis of the scattering noise image quality. Briefly introduces a method based on a variety of hardware and software to eliminate the scattering noise, and depth to eliminate beam filter array-based scattering method, the experimental results show that the method is simple and effective, with a high practical value. Image enhancement work, respectively, in the three aspects of the noise, contrast, and feature information. Integration method is more competitive than other methods eliminate the quantum fluctuation noise performance; decline in image contrast caused by the lens MTF adjusted based on energy image enhancement method to enhance global contrast and enhanced detail information; characteristics of the image in order to highlight proposed mask method based on local variance; Experimental results show that these three methods can effectively improve the image quality. Finally, the use of the above research results to practical applications in ray image quality, and a processing method, and lay a better foundation for its further recognition processing.