Machine vision is experiencing a sustained boom due to the requirements of new IIoT applications. It is well known that Industry 4.0 thrives on use case scenarios, and these in turn are spurred on by technological innovations. Matrix Vision, a brand of sensor and automation specialist Balluff, knows the requirements of plant and mechanical engineering inside out and is constantly developing high-performance hardware and software for industrial image processing.

The company is now expanding its portfolio with two 3D stereo cameras, one with 1.2 and one with 12 megapixel resolution. "Our camera systems help machine and plant manufacturers to avoid complex feeding kinematics when distributing individual parts from a large quantity, which are often technically complex and costly and usually require a lot of space," explains Fabian Furtner, the product manager responsible. "Bin picking is often used, in which an industrial gripper robot is guided by image processing to pick parts from a box. The use of camera systems for such Industry 4.0 applications reduces the technological effort, saves space and protects the environment."

It all depends on the perspective

The 3D cameras are based on stereo technology: images of a scene are captured slightly offset to each other via two camera lenses. The image processing goes far beyond that of conventional digital cameras, as depth information is also calculated as a third dimension. The smaller camera system from Matrix Vision has an "eye distance" of 65 mm, the larger one 160 mm. In addition, there is an Nvidia board that calculates a 3D image directly in the camera by correlating structures in the left and right camera image.

This technology is particularly interesting for logistics and robotics. In addition to the localization of objects in a planar (X-Y) surface, the Z-direction also plays a role, for example to determine the position of an object relative to the observer in order to control a gripping process. But what are the benefits of the ten times higher resolution of 4,096 x 3,008 pixels and a depth measurement range of 0.5 m to 4.0 m of the large system? "The high-resolution 3D camera system allows objects to be determined from a greater distance and much smaller objects to be recognized. You don't have to be so close to the scene with the camera, which is an advantage if there are unfavorable conditions such as high temperatures or if the robot needs to be given sufficient freedom of movement," explains Furtner.