3D Machine Vision Technology to Revolutionize the Vision Guided Robotic Industry
“blind” robots or traditional industrial robots are designed to perform a simple task, whereas vision guided robots with an advanced vision system are capable of performing critical tasks with variation and flexibility. In past few years, several techniques were introduced to automate the procedure of gripping parts of an industrial robot as a substitute to the existing manual part procurement. Due to the rapidly progressing machine vision technology, vision sensors these days are acting as a vital role in the 3D robot positioning systems. An embedded vision sensor robots have a greater consciousness of the scene. It can hold objects, which can be stacked, loosely located or non-fixture. Thus, it empowers the robot to hold/grip objects that are delivered in bins, racks, or on pallets. Regardless of the position, a vision-guided robot (VGR) can pinpoint an object for further processing. This generic solicitation of robotic guidance is smeared in industries such as automotive for the location of sheet metal body parts, power train components, complete car bodies, and other components used during assembly. Other industries such as pharmaceutical, food, daily products, and glass use vision-guided robotics as well. With respect to the industry needs, two significant practices have emerged: 3D and 2D machine vision systems.
The 2D machine vision is a well-applied technique and is implemented since the past few years. 2D vision system trace the object in three degrees of freedom (x, y, and roll angle) provided on a single image. Subsequently, the primary constraint of 2D vision is its incapability to identify part rotation outside of a single plane. However, the ineptness to identify a part rotation outside single plane does not suffice in many applications, for example, the precision fixtures in order to achieve greater versatility. 2D vision robotics are usually applicable in picking objects from conveyor belts. Calibration of such robotic systems requires relatively simple methods.
Whereas, the 3D machine vision robot can locate the parts in six degrees of freedom (x, y, z and yaw, pitch, roll). A 3D vision system computes the pose of the object analytically created on two overlapping images, or multi-vision systems, which conventionally combine the stereo-systems to increase precision and robustness. The 3D vision robotics help the robot to hold or handle a rigid object using information resulting only from one image. The distances between the object features have to be known to the system beforehand to compute the object pose relatively based on some predefined criteria. With the advent of 3D vision technology, which is achieved with the help of a 3D camera, laser and Time-of-Flight (ToF) camera, is unlocking newer potentials for robotic operation. Task such as object tracking, bin picking, and product profiling incurs the capability to generate three dimensional (X, Y and Z axis) image data, and the 3D image capturing technologies has revolutionized the robotics industry.
There are various advantages of 3D vision-guided technology, such as it provides more valuable data including information related to object flatness, surface angle and volume; measurement stability; precision and repeatability; multi-sensor stitching, and precision robotic guidance, etc. In addition, 3D vision-guided technology assist the robot in handling boxes of various dimensions for palletizing.
VGR are used across various verticals to concentrate on extracting, characterizing, and interpreting information of a particular product and perform desired autonomous task. The VGR companies (software and hardware providers) are constantly focusing on developing advanced technologies that would enhance the automation scenario to a greater extent and help the end-user industries to intensify their process.
Further, the VGR ecosystem has three major participants; the hardware provider (robots and camera), software provider, and service providers (including system integrators). 2D and 3D vision cameras are integrated with the vision system robots. 3D vision cameras are most common camera type as it offers a holistic view of the object upon which the action is executed. Whereas, the software providers offer essential software for the robots guidance. The working principle of these VGR is based on the performance of camera which captures the visual and sends it to the computer. On the other hand, service providers are system integrators who integrate the hardware and software into an existing or new industrial robots to enable them as VGR.
One of the significant hardware that transforms an industrial robot to vision-guided robots are cameras. The camera includes 2D camera and 3D camera. Sometimes it also consists of a time-of-flight camera. Majorly, vision-guided robots are a combination of 2D as well as a 3D camera. For instance, a 3D camera locates and picks random parts in a bin and then a 2D camera analyses the orientation of each piece and then places them on the conveyor belt. With the combination of laser 3D and ToF scanning camera sensors, some of the vision-guided robotics gain the resolution to operate in a broader spectrum of parts. A 3D ToF camera scans the time taken by a streak of laser light to travel the distance between the object and the camera. By this, it measures the actual distance, depth, and placement of the object, thus giving the ability/advantage to work on poor/any lighting conditions. With the improvements of 3D vision, the advancement of a random matching of a pattern has enabled new applications such as random bin picking and placing and quality control inspection.
Another critical component of VGR is their software. Software is mere algorithms that guide the robotic from colliding with the parts, bins, or from itself. The guidance software has to be robust in vision-guided robotics as each instance, the parts or bins require different path plans and intertwined. Vision guided robotics software is used across various verticals to concentrate on extracting, characterizing, and interpreting information of a particular product and perform desired autonomous task, whether mechanical or electrical. The software provides easy calibration as well as support for robots of different manufacturers, thus, it eliminates the apprehension of industry verticals to change to different brand of robots. Thus, it eliminates the apprehension of industry verticals to change to different brand of robots. Present-day software includes predefined settings such as grip correction, pick & place, as well as camera settings at par with the manufacturer brand.
Service providers can also be termed as system integrators, they integrate the vision system and software to an existing or new industrial robots. Service provider plays critical role in the deployment of the vision robotics as they offer a custom solution with a proven track record and provide post-installation service. The primary function of a service provider is to perform a feasibility study on the client’s requirement, identifying the correct vision system & software, providing cost-saving tips, training of vision robotics software usage, and incorporating the system into the factory.
Vision guided robotics industry is expected to evolve radically in the the automotive, electronics, and food processing industries in the near future. These industries are dominated by the 3D vision-guided technology, owing to the massive technology demand for applications such as arc welding, cutting, and palletizing. The latest technology developed by the robot manufactures are the 3D ToF vision-guided technology, which includes the integration of a 3D camera with ToF sensor, enabling the robots to perform tasks with higher accuracy and increasing the operational efficiency in any lighting condition.