Study Of Color Distortion Correction For Projection Display Image

As a type of large-field image display, the projection display, in comparison with other displays, exhibits a significant feature that the projection device is apart from its screen and thus own huge potential in the future development of display technology. Presently the projection display is drawing a wide attention owing to its superiority. The existed projectors could properly reproduce the color of the projected image on the uniform white diffuse screen in the dark environment. However, the projection display is required to be performed on the ordinary surface in our daily life with the extension of its application and its development towards mobile technology, which would result in the inferior quality of the projection display image due to the distortion of color reproduction. Herewith, it makes great importance for the development of projection display technology to research on the color distortion and its correction of the projection display image on the usual surface in our real world. In view of this purpose, the theoretical and technical study was explored in this dissertation for the color distortion correction of the projection display image on the diffuse reflection surface with light color patterns or textures.According to the principle of the projection display, the physical mechanism of the color reproduction on the colorized surface was analyzed and its physical model was established for the existed projection display devices. The distortion aspects were then physically and perceptively discussed for the projection display image on the colorized surface. Based on the projector-camera system (ProCams), the methodology was proposed for the color distortion correction of the projection display image, and the judgment rule was analyzed in consideration of the image sources, the characteristics of projection display on the colorized surface and the image reproduction intents. Thereafter, the geometric model and colorimetric model were developed for the transformation from the input pixel to output pixel, and the general mathematical description about the color distortion correction of projection display image was presented based on the pinhole model and color reproduction model.Taking the characteristics of the ProCams into account, the geometric mapping algorithm was proposed, including the procedures of sampling segmentation, region coding, region locating, region identification and pixel mapping, and the evaluation results demonstrated the mapping accuracy of pixel level. Meanwhile, the display characteristics of the projector, the response characteristics of the camera and the transmission characteristics of the ProCams were tested, and then the colorimetric characterization model for the dark and lighting environment were respectively constructed in the camera color space according to the specialties of projection display on the colorized surface in the different environments. Considering the characterization efficiency and the practicability of this system, the quasi-online and online colorimetric characterization algorithms for the individual environments were respectively proposed. Several ISO standard images were adopted as the test images to estimate the performance of the characterization algorithms, and the results indicated that there was no evident difference between the algorithm accuracies for the different environments and all satisfied the image correction demand, in which the mean error of the R, G, and B channels was about 10 for the forward prediction and about 20 for the backward prediction.Based on the current studies of the image quality assessment, the solution of the input compensation image was separated into two processes, i.e. the calculation of the objective output color and then the prediction of the input compensation color. Meanwhile, based on the color reproduction evaluation and its intents, the calculation methods of the objective output color for the dark and lighting environments were respectively developed depending on the objects of colorimetric and equivalent reproduction. Thereafter, the adaptive algorithms of the objective color and compensation color were respectively proposed for the individual environments and reproduction intents on the basis of the colorimetric characterization. Finally, the general framework for the color distortion correction of projection display image was constructed. For assessment of the proposed algorithms, the ISO standard images were employed to carry out the color correction experiments on the "patch" and "flower" colorized surfaces in the dark and lighting environments, respectively. The experimental results indicated that the corrected output images obviously improved the quality of the projection display image, which demonstrated the validity of the developed models and methods in the study of this dissertation.