为了使救援机器人在危楼环境下能够精确地检测人体关节点，以代替救援人员进入危险现场，提出了一种基于救援机器人的人体关节点检测方法。首先，通过Kinect相机获取人体关节点检测数据，采用归一化的方法对关节点数据进行平滑预处理。其次，对预处理数据加入人体关节长度约束和关节旋转角约束，从而生成每个人体关节点空间位置。最后，采用Tsai方法进行手眼标定，得到机械臂和伤员之间的距离。实验结果表明该方法在危楼救援环境下能够准确测量伤员关节点位置，验证了所提出检测方法的有效性。

In order to enable rescue robots to accurately detect human joint points in dangerous building environments to replace rescuers into dangerous scenes. A machine learning based human joint point detection method was proposed. Firstly, human joint detection data was obtained through Kinect cameras, and normalized methods were used to smooth and preprocess the joint data. Secondly, human joint length constraints and joint rotation angle constraints were added to the preprocessed data to generate the spatial positions of each human joint point. Finally, the Tsai method was used for hand eye calibration to obtain the distance between the robotic arm and the injured person. The experimental results show that the method accurately measures the position of the casualty joint point in the rescue environment of a dangerous building, which verifies the effectiveness of the proposed detection method.