- Why ZIVID
Mount the Zivid machine vision camera on the robot to confidently tackle your assembly tasks.
Point precision of better than 55 um means even small objects and their detail are easily detected and understood by your robot.
Automated assembly has unique and challenging demands for robotics. The tasks involve human levels of dexterity where parts must be singulated and grasped. These fine motor skills are reliant on the performance and quality of the 3D vision system. High precision is important, but of most importance is very high degree trueness, meaning seeing things as they really are, and where they really are.
Assembly tasks often require an intermediate re-alignment stage between picking up the object and the assembly operation where the object is placed into a jig to ensure a precise grasp. This is costly in terms of additional equipment, and space and introduces a time penalty to the operation. Zivid 2+ is ideal for robot-mounted operation, where the 3D vision can move with the robot to see the scene from any angle and position.
Being able to see objects and their assembly counterparts with a very high degree of trueness is essential for consistently reliable assembly tasks. The alignment processes are typically at a millimeter level. Without this ‘true-to-reality’ vision mispicks and misaligned assembly is highly likely.
Assembly takes place across a broad range of industries. There can be lots of other things happening from shutter doors opening to mobile machines moving around. All these other things can create temperature variations, vibration, and even knocks. A true industrial robot vision camera designed and tested to exacting standards is essential for reliable performance and repeatability.
Zivid Two as a robot-mounted 3D color sensor is inherently ideal for robot-guided assembly applications. Built to true industrial quality standards to survive in the toughest conditions, it delivers consistently high-quality 3D point clouds offering detail at 55 um point precision and true-to-life form and representation with < 0.2% dimensional trueness error.