为解决室外环境下基于可见光相机图像分析算法及应用系统环境适应能力差、易失效的现状，提出一种基于成像环境自感知的室外环境下可见光相机图像增强方法，以期提高可见光相机的输出效果，并为后续算法设计提供计算基准。首先，采用盲图像质量评估技术进行成像环境自感知模型建模，综合利用人机工效学实验、图像处理与模式识别理论，实现上述盲图像质量评估参数与环境自感知模型的设计；其次，在充分研究多种典型图像增强算法环境适应性与算法计算效果控制规律前提下，采用模式识别关联分析技术，设计出一种根据成像环境自感知评估结果控制的改进型图像增强算法；最后，利用一种无人机航拍系统，设计多种实验方法与评价指标，对图像增强算法的环境适应性进行评估与实证研究。通过课题研究，可以改善室外复杂环境光条件下可见光相机图像输出的质量，为后续高层算法的设计提供计算基准和设计依据，也为进一步了解可见光相机在复杂光照条件下的成像规律提供参考。

To change the phenomenon of poor environment adaption and function losing issue of the outdoor visible light camera based image analysis algorithms and application systems, a novel outdoor visible light camera image enhancement method, which uses the imaging environment self-perception technique, is proposed. This algorithm is designed for improving the output effect of the visible light camera and providing the computational reference for the following algorithms. First, the blind image quality evaluation technique is utilized to build the imaging environment self-perception model. The experiment method of man-machine ergonomics, together with the image processing and pattern recognition theory, are all employed to build the imaging environment self-perception model. Second, after the comprehensive researches on the environment adaption ability and the control law of the typical image enhancement algorithms, the relationship analysis technique of pattern recognition method is used to design a kind of improved image enhancement algorithm, which can tune the control parameters according to the evaluation result of the imaging environment self-perception model automatically. Finally, an Unmanned Aerial Vehicle (UAV) system is utilized to test the environment adaption of the improved image enhancement algorithm. A series of experiment methods and evaluation indexes are designed. By the research of our proposed subject, the output quality of the outdoor visible camera, which works under the complex environment light, can be improved. And the corresponding results can also provide the computational references and design guidance to the following high level algorithms. In addition, this research work will also get some knowledge about the imaging law of the visible light camera which works under the complex environment light.