针对移动服务机器人室内作业环境中场景物体在图像中表现为不同色块聚集的特点,提出一种基于色彩分层及多尺度滤波的场景分割方法.通过构建色彩分层模型对场景图像进行分层处理,设计算法对图层中连通域的数量和尺寸进行检测,根据检测结果设计多尺度滤波器进行目标分割.建立系统对机器人作业过程的全局环境及局部场景进行实验,场景分层和分割的平均正确率分别达到了97.1%和93.05%,实验结果表明,该方法能有效地对场景中具有明显色彩的区域进行分割.%The obj ect features gathered as different color blocks in the scene images for the indoor mobile robots,consequently,a kind of scene segmentation method was proposed in light of color layering and multi-scale filtering.Through constructing the color-layering model,the different image layers were got,the size and quantity of connected domains in different layers were detected,and according to the detection results,the multi-scale filters were provided for obj ect segmentation.Some segmenting experiments were done with the global environment and local scene for the robot in completing tasks,and the average accuracy of scene layering and segmenting are 97.1% and 93.05% respectively,the results show that the proposed method can effectively segment the scenes with salient color features.
展开▼
