Abstract: For dynamic environments including moving objects，a salient features extraction approach based on the thoughts of filtering those moving objects is presented.Firstly the local salient image features based scene recognition approach is introduced.Then the framework of salient features extraction with filtering moving objects is proposed.The implement of moving objects detection is discussed in detail.Experiments and analysis show that the approach can filter effectively moving objects in the environments.And the accuracy of scene recognition is improved.

Key words: dynamic environments, scene recognition, filtering moving objects, visual salient features

摘要： 针对存在运动目标的动态环境，提出一种基于动态目标滤除思想的显著特征提取方法，据此实现基于局部图像特征的场景识别。首先简要介绍基于局部显著图像特征的场景识别方法，然后提出了带动态目标滤除思想的显著特征提取框架，并详细讨论了运动目标检测及提取的实现。实验结果和分析表明，该方法能够有效地过滤环境中的运动目标，提高场景识别的精度。

关键词: 动态环境, 场景识别, 运动目标滤除, 视觉显著特征