中文摘要：

为实现投影仪在任意表面上的自适应投影，提出一种可以在日常环境光照条件下进行自动几何校正和色彩补偿的投影图像校正算法．该算法使用基于二进制编码结构光的离散映射集合方法，通过计算对应像素映射关系消除投影图像的几何畸变；然后应用一种优化的朗伯特反射模型构造投影表面的纹理空间，对原始投影图像进行预处理和全局亮度补偿，最终实现投影图像的颜色补偿．实验结果表明，基于文中算法在复杂环境中进行投影，可以在不规则几何表面上投影出符合视觉期望的无变形图像，并能够有效地消除投影表面固有纹理和环境光照对投影图像的干扰；将自适应复杂环境的投影图像校正算法应用于传统投影机摆脱了投影机对投影幕布的依赖，可以在任意环境中实现近似于白色平板幕布的投影效果．

英文摘要：

This paper proposes an adaptive calibration algorithm for projected images which enables ordinary projectors to project on arbitrary surfaces in daily environment. Firstly, the algorithm uses a discrete mapping method based on binary-coded structured light to register geometry of projection surfaces. Then an optimized Lambertian reflection model is applied to construct texture space of projection surfaces and environment illumination. Finally, based on the established model the original projection image is warped in geometry space and compensated in color space to make it shown as desired on the projection surface. Experimental results show that this algorithm can adapt to irregular geometric shape and complex inherent texture of projection surfaces in common ambient light environment and get projection effect on arbitrary surfaces as like to projection on uniformly white flat screens.