Currently, the main applications of machine vision are quality inspection, dimensional measurement, defect inspection, identification, and localization, while its application in robots is primarily for guidance and localization. Robot vision guidance systems are evolving from monocular vision to multi-view 3D vision guidance systems based on multiple lenses.
The emergence of different vision systems and robot controllers has made communication easier to integrate. The tighter the integration between the robot and the vision system, the easier it is for designers to link the three related coordinate systems (robot, vision, and the real world).
Leaving aside the applications of machine vision in other automation fields, the application of machine vision technology in robot vision systems alone has a very broad scope, which is mainly due to the development plan of my country's robot industry.
Machine vision will be widely used in the field of industrial robots in the future, and it mainly has four functions:
1. Guiding and Positioning: Visual positioning requires machine vision systems to quickly and accurately locate the part being measured and confirm its position. Machine vision is used for positioning during loading and unloading, guiding the robotic arm to accurately grasp the part. In the semiconductor packaging field, equipment needs to adjust the pick-up head based on the chip position information obtained from machine vision to accurately pick up and bind the chip. This is the most basic application of visual positioning in the machine vision industry.
2. Appearance Inspection: This step involves inspecting products on the production line for quality issues and replaces the most manual labor. In the pharmaceutical field, machine vision primarily inspects dimensions, bottle appearance defects, bottle shoulder defects, and bottle mouth defects.
3. High-precision testing: Some products have high precision, reaching 0.01~0.02m or even micrometer level, which cannot be detected by the human eye and must be completed by machine.
4. Recognition involves using machine vision to process, analyze, and understand images to identify targets and objects of various patterns. It enables data traceability and collection, and is widely used in automotive parts, food, and pharmaceuticals.
To enable robots to perform more complex tasks, they need not only better control systems but also greater sensitivity to environmental changes. Robot vision, with its large information capacity and completeness, has become the most important robot perception function. It excels particularly in 3D visual guidance, making it the industry's most advanced visual guidance product, capable of covering a wider range of robotic applications.
So, what are the current development directions for machine vision and robotics?
First, a large number of user-friendly and fully localized Chinese software platforms have emerged in the field of machine vision software.
Second, visual light source products based on visual sensors or visual lenses do not require a separate power supply.
III. 3D vision inspection and robot 3D vision guidance systems are highly recommended.
Fourth, the issue of open source for the robot system ROS is mentioned.
V. Localization and autonomous navigation issues in robot integrated applications.
VI. Losses that may be caused to enterprises by industrial shutdowns and accidents.
It is evident that the integration of machine vision and robotics primarily addresses the localization problem. High open-source nature, high reliability, and high ease of use have become fundamental requirements for related automation products. Overall, related trends include the miniaturization and high integration of vision products themselves, the development of 3D vision, the realization of autonomous robot navigation, and integration with big data and intelligent control systems. The future is promising!
