Rechearch Of The Digital Cameras Colorimetric Characterization

Digital cameras are widely used in many fields. Nevertheless, in the process of taking pictures the final pictures always appear in different colors on CRT-display and liquid crystal display and printer such color-display equipments because it doesn't feature those original colors on the equipments. This kind of digital cameras which has no function to check colors cannot satisfy the demand for color-copying and transmission nowadays. Hence, it is a desperate requirement to study colorimetric characterization of digital cameras for the modern application.According to the demand for the consistency and stability in the color-transmission process when using digital cameras, this paper focuses on the basic research about the technique of colorimetric characterization of digital cameras by using method of artificially neural networks. First, the author analyzes the causes of color-deviation in terms of both software and hardware when using digital cameras. By using comparisons between different ways of colorimetric characterization of several famous cameras in our country and cameras abroad, the author points out the direction of tackling the problem. Then, the author puts his emphasis on the features of artificially neural networks in the aspects of nonlinear code-reflection, auto-check and deviation reversal transmission and then proposes a way of colorimetric characterization of digital cameras based on BP neural networks. Finally, the author uses three digital cameras as the objectives to carry out the design, establishment and check of colorimetric characterization and finishes the choice of hardware and the building of the experimental platform. The programming realizes the transformation from RGB colorimetric space related to equipment to Lab colorimetric space not related to equipment. It then analyzes the exercise and check deviations derived from colorimetric space transformation model based on BP neural networks.The result shows that the study methods about colorimetric characterization based on BP neural networks basically satisfies the demand of changing from RGB colorimetric space related to equipment to Lab colorimetric space not related to equipment. This system has stability and convenience in operation. It has the value to be applied.