Technology of real scene restoration

Through the real scene restoration algorithm, the real object is modeled. The color depth camera and depth camera are used to scan around the object in all directions to get the original physical characteristic data of the object, the object can be scaled in virtual space (displayed by computer) and the object features (real size, color, holes, edges, edges, and positions in space, etc.) can be preserved. In order to observe the effect of robot simulation in virtual space, such as painting, polishing, assembling, etc.

Image processing and deep learning techniques

Pre-processing based on image and depth learning identifies the body joints through point cloud analysis and image processing, and automatically divides and classifies the corresponding regions of the body according to the features of the seam contour and depth learning

Automatic Planning Technology based on scanning model

For a single detection area, such as a car door, a hybrid sampling measurement strategy is used for point planning. That is, according to the number of detection points, the detection area is divided into several subspaces, and then an algorithm is applied to each subspace to select specific points as detection points or path points.

The trajectory planning curve can be formed after the selected points are connected in order, and the trajectory curve can be transformed into the motion information of the manipulator by the Motion Planning Algorithm. Once the robot arm has the information, it can complete the task efficiently.

Intelligent planning and active obstacle avoidance technology

The motion trajectory of the robot can be predicted in advance by simulating operation, and the real spatial relationship of objects such as object, robot and so on can be obtained by 3D solid space modeling. In order to achieve automatic shielding, collision detection. When a foreign object is placed in the working area of the robot and interference is detected in the simulated operation, the robot is automatically alarmed and prevented from operating, pause the robot. Through collision detection and risk early warning mechanism to prevent accidents, thereby protecting the personal safety of workers and enterprise plant property.

Intelligent non-teaching technology

At present, the robot application market is mainly based on manual teaching and guiding robot. It takes several days to teach the robot in a conventional manual instruction painting, and can only teach the object it is supposed to teach. The manual teaching time is long, the applicability is bad, has the security hidden danger. The motion trajectory generated by simulation is used to ensure real-time, high efficiency, stability, safety and reliability. Through the big data and the corresponding algorithm developed, the computer intelligent planning robot spray path. Develop Special Professional Remote Control System (add new PCL control unit, develop new software control algorithm and input robot system, etc.) to operate the robot, thus achieve expert remote assistance.
Cooperative Robot
Harmo College
Intelligent Equipment
Application Case
Cooperative Robot
End Actuator
Software Version
Service Aupport
Common Problem
Data Download
About Harmo
About Harmo
Join Harmo
Sales Network
Contact Us
About us
  1. +86 027-5976 0630
  2. 湖北省武汉市洪山区光谷大道62号光谷总部国际9栋902
  3. 湖北省黄石市经济开发区大棋路88号黄金山产业配套园15-16栋
  4. info@hmbots.com