摘要
虽然采用穿戴式传感器进行坐姿检测的测量结果准确,但是会对被测试者造成束缚并且成本较高。提出使用Kinect的骨骼角度测量算法对坐姿进行非接触式的实时检测并判断坐姿意图。通过单片机来控制座椅从而适应用户当前坐姿;通过骨骼角度测量算法测量头部和背部角度,从而确定坐姿的意图;利用信心检验方法排除冗余数据,提高数据测量的准确性;将Kinect检测数据与彩色摄像头录像进行对比,Kinect的测量准确性与稳定性均在可接受的范围之内。原型机测试结果表明,该方法可以较为准确地判断出用户的坐姿意图,并能准确控制自动调节椅来适应不同坐姿行为,到达矫正用户坐姿的目的。
Although the measurement results of the sitting position detection using the wearable sensor are accurate, the testers will be fettered and the cost will be high. Therefore, we used kinect-based bone angle measurement algorithm to detect the contactless real-time sitting posture and judged the sitting intention. The single chip microcomputer was used to control the seat to suit the user＇s sitting position. The angle of the head and back was measured by the skeletal angle measurement algorithm to determine the intent posting. The method of confidence test was used to eliminate the redundant data, which improved the accuracy of data measurement. The kinect detection data were compared with color camera video recording, and the measurement accuracy and stability of kinect were acceptable. The prototype test results show that the it can accurately determine the user＇s intention to sit, and can automatically control the automatic adjustment chair to adapt to different sitting behavior, to achieve the propose of correcting the user＇s intention to sit.
作者
吴剑锋
马梦鑫
蒋濛婷
罗凯
Wu Jianfeng;Ma Mengxin;Jiang Mengting;Luo Kai(Zhejiang University of Technology,Hangzhou 310023,Zhejiang,China)
出处
《计算机应用与软件》
北大核心
2018年第10期194-199,共6页
Computer Applications and Software
基金
国家自然科学基金项目(51375450)
关键词
骨骼角度测量
KINECT
坐姿意图
坐姿矫正
Skeletal angle measurement
Kinect
Sitting posture intention
Sitting posture correction
