【摘要】 根据心理物理学、神经生理学、认知神经学等学科在视觉认知领域的部分研究成果，结合机器视觉、图像处理领域在图像增强方面已经提出的一些方法，提出了结合先验知识的多窗口结构下的分块中值滤波方法，在每一个窗口内单独进行处理与分析，突出了视觉处理目的，减少了运算量， 节省数据存储空间，达到了令人满意的滤波效果，能够在原始图像比较复杂的情况下，较好地对其进行预处理，可以改善、提高后期图像处理过程，如图像分割、图像分析的正确性和有效性。更多还原

【Abstract】 Based on the research progress in the vision and cognition field of psychophysics, neurophysiology and cognitive neuroscience, combined with the methods proposed in the image enhancement in digital image processing field, this paper brings forward a self-adapting median filtering method under muti-window structures in the digital image processing. It does the data processing and analyzing in seperated windows, which stresses the object of the vision process, reduces the time of computation and the space of the data memory. It is a useful and efficient image preprocessing methods when the original image are not satisfying.更多还原