The paper deals with the problem of object recognition for the needs of mobile robotic systems (MRS). The emphasis was placed on the segmentation of an in-depth image and noise filtration. MS Kinect was used to evaluate the potential of object location taking advantage of the in-depth image. This tool, being an affordable alternative to expensive devices based on 3D laser scanning, was deployed in series of experiments focused on object location in its field of vision. In our case, balls with fixed diameter were used as objects for 3D location.
展开▼
机译：本文针对移动机器人系统（MRS）的需求处理对象识别问题。重点放在深度图像的分割和噪声过滤上。 MS Kinect用于利用深度图像评估对象定位的潜力。该工具是基于3D激光扫描的昂贵设备的经济实惠替代品，已针对其视野中的对象定位进行了一系列实验。在我们的案例中，使用固定直径的球作为3D定位的对象。
展开▼
