Drives & Controls April 2024

40 n MACHINE VISION April 2024 Four eyes guide container-loading operation The German machine vision specialist VMT Vision Machine Technic Bildverarbeitungssysteme – part of the Pepperl+Fuchs group – has developed the robot-based system that uses four 3D cameras for automated loading and unloading of pressed parts to and from large open-sided containers. The system, called FrameSense, was developed for a carmaker that wanted to improve the cycle times on its press lines. The work ow the system is designed to automate is part of many manufacturing operations. A component comes out of a machine – in this case, a press – and moves along a conveyor belt to a container, where it is stacked. When the container is full, it is moved to the next production step, such as assembly into a vehicle. Until now, employees have been responsible for loading and unloading the containers. This apparently simple task is actually more complex than you might think. Before the insertion process, the rst step is to nd free space for the part in the container. Any interfering factors, such as interlocks, must be dealt with and the “load box” needs to be checked for any defects. These tasks are now all handled by a robot combined with a vision system. This is a technological challenge because the containers come from diˆerent suppliers, and can vary in their dimensions. Four cameras For automated loading and unloading, the position of several features of the containers must be determined to allow “multi-vector” correction of the robot path. The type, shape and position of each container must be checked to ensure reliable, collision-free guidance of the robot. This capability has to be integrated into the existing production process. Time delays must be eliminated and the positioning of the components must be accurate to within millimetres. To achieve this VMT uses four 3D cameras, each recording part of the image eld. This can consist of two containers, each measuring around 1.5 x 2 x 1.5m. Two of the cameras focus on each container. This produces data from two perspectives for a high-quality 3D point cloud. The point clouds of all four sensors are combined for the subsequent evaluation. Relevant features of the container are registered “regions of interest” (ROIs) in the point cloud. A registration is the exact position of a feature using a model with six degrees of freedom. Interference contours are searched for in other ROIs which could lead to collisions during loading. Finally, the complete picture is compared with a stored reference model. In this way, the containers can be checked simultaneously and automatically for their condition and position. Even deformed or slanted containers can be processed. This information is also recorded for use in a quality management system where the condition of all containers can be traced. The calibration as well as the consolidation of the measurement data and their subsequent evaluation are carried out by a separate industrial PC with visualisation, operating elements and connection to the robot controller. The main result of the image processing is the multi-vector correction. The robot is adjusted to insert the component at the next suitable location. Secondary results are error messages caused by interfering edges or objects in the container that would prevent lling. The data can detect damaged containers. The image processing takes place in VMT’s MSS (Multi Sensor Systems) imageprocessing software. VMT is using 5-Megapixel IDS Ensenso C 3D cameras, supported by high-intensity projectors, which allow large volumes to be measured. This is important for the FrameSense application, because the robot can only reach the containers to be lled up to a certain distance. By combining the 3D container inspection with automated loading and unloading, the technology can help to automate previously manual processes. Against the background of a shortage of skilled workers, this can make an important contribution to automation in the automotive industry, and others. VMT product manager Lukas Neumann believes that that 3D vision systems with high resolutions and powerful projectors will be particularly attractive for intralogistics applications where “high-precision components have to be gripped from a great distance with a large measuring volume”. He also foresees similar cameras with high projector power but lower resolutions and fast recording being used for other stacking or bin-picking applications. n A German vision specialist has developed a technology that combines robots with four 3D cameras to load and unload pressed parts automatically from large containers with di ering shapes. Robots combined with overhead 3D cameras and powerful projectors are being used to load and unload pressed parts into containers of varying shapes and sizes