论文部分内容阅读
家庭日常工具的部件功用性主动认知是家庭服务机器人智能提升的重要方面。为满足服务机器人实时自主作业的需要,提出了一种基于结构随机森林(SRF)的工具部件功用性快速检测算法。在离线训练阶段,利用SRF训练功用性边缘检测器与功用性检测器,并通过评估功用性检测结果的Fβ值确定工具各部件功用性对应的先粗糙后逐步精细化(coarse-to-fine)阈值。在线检测阶段,首先使用功用性边缘检测器计算功用性区域边缘的初步概率图,继而加以coarse-to-fine阈值滤波得到包含工具部件功用性的外接矩形区域,最后对该区域使用功用性检测器进行检测。实验结果表明,在普通非图形处理器系统下,相较于现有的全局搜索检测方法,本文方法对各功用性部件的检测效率均明显提升,且召回率和精度都有提高。
The functional awareness of the components of everyday household tools is an important aspect of intelligent enhancement of home service robots. In order to meet the need of real-time autonomous robot for service robots, a rapid detection algorithm of tool components based on structural random forest (SRF) is proposed. In the off-line training phase, the SRF is used to train the functional edge detector and the functional detector, and the coarse-to-fine coarse-to-fine correspondence is determined by evaluating the Fβ values of the functional test results. Threshold. In the on-line inspection phase, the preliminary probability map of the edge of the functional area is first calculated using a functional edge detector, followed by coarse-to-fine thresholding to obtain a circumscribed rectangular area containing the functional parts of the tool. Finally, a functional detector Test. The experimental results show that, compared with the existing global search and detection methods, the detection efficiency of each functional component is significantly improved, and the recall rate and accuracy are improved under the conventional non-graphics processor system.