概率图模型及其应用作为一个具有挑战性的研究领域目前已成为一个新的研究热点。概率图模型为解决智能信息领域的不确定性问题提供了重要途径。尽管目前概率图模型还处于不断发展之中，但近年来基于概率图模型的图像和视频智能信息处理的应用研究受到人们的关注，出现了许多有效的算法，这些算法为解决一些传统的图像和视频智能信息处理问题提供了新的途径。本文首先对概率图模型的3种重要表现形式、特性和主要技术进行了分析和讨论，在此基础上，以概率图模型在图像和视频中的应用为线索，对目前基于概率图模型的图像和视频智能信息处理的主要技术进行了概述和比较研究；最后对概率图模型所存在的一些问题及进一步的发展进行了展望。

Probabilistic graphical model(PGM) and its application have become a new hot-spot as a challenging research. It provides an important means for resolving the uncertainty of intelligent information field. Although PGM is still in its development, the application of intelligent information processing of images and videos based on PGM has emerged and many effective algorithms has emerged which provide a new strategy to solve for the traditional problems. Based on the three important expressions, the characteristics and the main technology of PGM,we analyze and introduce the primary technology of intelligent information processing of images and videos based on PGM. Finally, we discuss the disadvantages and development trend of PGM.