Machine Vision
Performing repetitive motions is not difficult for industrial robots, but in an unstructured environment, robots must be able to recognize, analyze, and interpret their surroundings.
Principle
A technology that captures images of the external environment through sensors such as industrial cameras, then processes them with a processor or computer so the robot can recognize, understand, and respond accordingly. It integrates knowledge from optics, mechanics, electronics, and computer hardware and software, making it an important branch of modern technology.
Use Cases
- High-speed sorting
Enables intelligent sorting of goods on high-speed conveyor lines: the vision system streams workpiece position and orientation data, guiding robots to complete high-speed sorting independently or collaboratively. - Automatic palletizing and depalletizing
By combining vision with the NRC system's palletizing process, automatic depalletizing and palletizing can be achieved without complex robot programming. - Loading and unloading
Applicable to machine loading and unloading in stamping, CNC, bending, and other scenarios. With 3D vision, the system automatically identifies workpiece positions and orientations in the bin, ensuring accurate loading and unloading every time — no operator supervision required.
Advantages
- Precise positioning: accurately locates each workpiece's spatial coordinates and orientation, so incoming parts can be grasped without being arranged or pre-positioned.
- Easy deployment: no complex mechanical fixturing required, with the reliability of automated machinery.
- Simple to use with flexible parameter configuration: users can customize vision parameters, with an intuitive, clean interface.
- Conveyor tracking: iNexBot's proprietary tracking process effectively reduces tracking errors without complex programming.
- Rich command set: the vision enable, trigger, counting, and running processes are broken into individual commands, letting users freely customize their own vision processes.
- Multi-process integration: vision processes can be combined with other processes such as loading/unloading and palletizing, reducing manual teaching time and improving efficiency.
- Multi-robot collaboration: the vision system automatically assigns multiple picking tasks to several robots on the line for collaborative grasping, achieving 1+1>2 efficiency.
- Multi-brand support: compatible with cameras from Hikvision, Sensopart, Leica, Cognex, Keyence, and other brands
Future Trends
- Continuously advancing intelligence: as deep learning and other technologies evolve, robot vision systems will become increasingly intelligent, recognizing and analyzing image information more accurately with greater autonomy and adaptability.
- 3D vision technology: 3D vision captures the three-dimensional spatial information of objects, which is essential for tasks such as robot navigation and manipulation. 3D vision will see widespread adoption in robot vision systems in the future.
- Miniaturized, lightweight hardware: as microelectronics and new materials advance, the hardware of robot vision systems will become increasingly miniaturized and lightweight, improving robot portability and adaptability.
- Multi-sensor fusion: multi-sensor fusion processes information from different sensors together, improving the perception and robustness of robot vision systems. This technology will be widely adopted in the future.
