Detailed Explanation of Robot Vision Sensors
A new type of industrial robot has reached the forefront of the industry, its main feature being its ability to safely assist humans in work. Many people talk about them online, but have you really looked into them?
In 2008, many people loved them simply out of curiosity. In 2012, robots were considered fashionable. But in 2013, a large number of competitors began their robot era and entered the artificial intelligence battlefield.
Robots have become one of the hottest topics and a focus of attention for many media outlets. Many now say that robots are our future. A series of robots are entering the manufacturing industry.
Regarding robot vision sensors
Sensing technology is one of the three essential elements of modern robots (perception, decision-making, and action). Industrial robots are equipped with different types and specifications of sensors depending on the tasks they perform. Generally, based on their application, robot sensors can be divided into two main categories: internal sensors used to detect the robot's own state and external sensors used to detect relevant environmental parameters.
Today's topic is visual sensors, one of the external sensors for robots, which are essentially the robot's "glasses."
Robots having "eyes" is extremely important to us. Humans obtain most of their information from the outside world through their eyes; the number of visual cells in humans is more than three thousand times that of auditory cells and more than one hundred times that of sensory cells in the skin. If we want to endow robots with higher levels of intelligence, they must acquire more information about their surroundings through their visual systems.
Example: Typical applications of robot vision:
Welding robots use vision systems for job positioning
A vision system guides a robot to perform a painting operation.
The handling robot uses a vision system to guide the electromagnetic chuck to grasp the workpiece.
Image: Robot welding
A robot's vision must be able to understand information in three-dimensional space. In other words, a robot's vision is different from text recognition or image recognition and requires three-dimensional image processing.
Because visual sensors can only obtain two-dimensional images, the images obtained from viewing the same object from different angles will also be different; the brightness and distribution of the images will also be different depending on the position of the light source.
To address this issue, many measures have been taken, including improving external environmental conditions, enhancing the functionality of the visual system itself, and employing better methods for information processing to reduce the burden on the visual system.
1. Video camera
2. Optoelectronic conversion devices
(CCD sensor, MOS image sensor)
3. PSD sensor
4. Shape recognition sensor
5. Industrial robot vision system
(Basic principles of industrial robot vision systems, and industrial robot systems that use vision recognition to grasp workpieces)
The robot's "eyes" don't actually "see" anything. Instead, the robot has embedded visual sensors that scan its surroundings to detect obstacles. When it encounters a human or an obstacle, the robot can slow down or stop.
As more and more robots enter factory floors, their safety remains a major concern. Robotics and safety are inextricably linked. You cannot have human-robot collaboration without mitigating the risk of injury.
If your robot is handling sharp objects, it is unsafe for a person to be nearby without protective safety measures. Another scenario is if the robot is handling heavy objects, which could cause injury if they fall or become projectiles at a certain speed.
In order for people to feel safe around these robots, we need to understand what the robots will do next.
Low cost, automation is true robotics. By adding robots, factories are able to achieve higher output with the same number of people, reaching maximum production efficiency at the lowest cost.
By combining robotics with autonomous mobile robots, augmented reality, wearable devices, and other advanced technologies to equip smart, digital factories, you have an entertaining, forward-looking future of automated manufacturing.
In the future, robots will work hand in hand with human teams to improve efficiency and productivity.
Technology and applications of intelligent vision sensors
Visual sensing technology is one of the seven major categories of sensing technology. A visual sensor is a sensor that calculates the characteristic quantities (area, center of gravity, length, position, etc.) of an object by processing images captured by a camera, and outputs data and judgment results.
I. Overview of Visual Sensors
Visual sensing technology is one of the seven major categories of sensing technology. A visual sensor is a sensor that calculates the characteristic quantities (area, center of gravity, length, position, etc.) of an object by processing images captured by a camera, and outputs data and judgment results.
II. Classification
1. 3D visual sensing technology
3D vision sensors have a wide range of applications, including multimedia mobile phones, webcams, digital cameras, robot vision navigation, automotive safety systems, biomedical pixel analysis, human-machine interfaces, virtual reality, surveillance, industrial inspection, wireless long-range sensing, microscopy, astronomical observation, marine autonomous navigation, and scientific instruments. All these diverse applications are based on 3D vision image sensor technology. In particular, 3D imaging technology has urgent applications in industrial control and autonomous vehicle navigation.
2. Intelligent visual sensing technology
Intelligent vision sensors, also known as intelligent cameras, are one of the fastest-growing new technologies in the field of machine vision in recent years. An intelligent camera is a small machine vision system that integrates image acquisition, image processing, and information transmission functions; it is a type of embedded computer vision system. It integrates an image sensor, digital processor, communication module, and other peripherals into a single camera. This integrated design reduces system complexity and improves reliability. At the same time, the system size is significantly reduced, broadening the application areas of vision technology.
The ease of learning, use, maintenance, and installation of intelligent vision sensors, along with their ability to quickly build reliable and effective vision inspection systems, has led to the rapid development of this technology.
III. The Foundation for the Implementation of Visual Sensing Technology
The image acquisition unit of a vision sensor mainly consists of a CCD/CMOS camera, an optical system, an illumination system, and an image acquisition card. It converts optical images into digital images and transmits them to the image processing unit. Commonly used image sensors are mainly CCD image sensors and CMOS image sensors. The implementation principles, advantages, and disadvantages of these two types of sensors will be introduced below.
IV. Applications of Visual Sensing Technology
1. Automotive body visual inspection system
Body forming is one of the key processes in automobile manufacturing, with stringent requirements for various body specifications, necessitating 100% inspection of the body. Traditional body inspection methods utilize coordinate measuring machines (CMMs), which are complex to operate, slow, and time-consuming, allowing only for random sampling.
Typically, the key dimensions of a vehicle body are mainly the windshield size, the edge positions of door mounting locations, and the positions of positioning holes. Therefore, vision sensors are distributed near these locations to measure the spatial dimensions of the corresponding edges, holes, and surfaces. A measurement station is designed on the production line. After the vehicle body is positioned, it is placed within a frame composed of longitudinally and transversely distributed metal columns and rods. Vision sensors can be flexibly installed on the frame as needed. The number of vision sensors installed depends on the number of measurement points (typically each vision sensor measures one measurement point). Different types of sensors are used, including binocular stereo vision sensors and contour sensors.
The measurement system works as follows: the car body is transported from the production line to the measurement station for accurate positioning. Then, the sensors start working in the required sequence. The computer collects and processes the images of the detection points, calculates the spatial three-dimensional coordinates of the measured points, compares the calculated values with the standard values, obtains the detection results, and then sends the car body out of the measurement station.
2. Online visual measurement system for steel pipe straightness and cross-sectional dimensions
Seamless steel pipes are an important industrial product in industrial production, and their quality parameters are crucial manufacturing data. Among them, the straightness and cross-sectional area of the steel pipe are the main geometric parameters and are key to controlling the manufacturing quality of seamless steel pipes. However, the measurement of these parameters is difficult due to the following reasons: (1) Seamless steel pipes are measured using non-contact methods, and the manufacturing environment is harsh; (2) Seamless steel pipes have large spatial dimensions, which also requires the detection system to have a large measurement space. The emergence of visual sensing technology has solved the above problems. Visual sensing technology uses non-contact measurement and has a large measurement range.
The measurement system consists of multiple structured light sensors. The light plane projected by the structured light projector on the sensor intersects with the steel pipe being measured, obtaining a partial arc on the circumference of the steel pipe cross-section. The sensor measures the position of this partial arc in space. Each sensor in the system measures a partial arc on a cross-section. Through appropriate mathematical methods, the cross-sectional dimensions and the spatial position of the cross-sectional center are obtained by arc fitting. The straightness parameters are obtained from the spatial envelope of the cross-sectional center distribution. Under computer control, the measurement system can complete the measurement within seconds, meeting real-time requirements.
3. Three-dimensional shape visual measurement
3D topographic digital measurement technology is a fundamental supporting technology for reverse engineering and digital product design, management, and manufacturing. Its mechanism for achieving 3D topographic digital measurement combines non-contact visual measurement with the latest high-resolution digital imaging technology. Since the objects being measured are often large and have complex surfaces, the measurement process typically consists of two parts: local 3D information acquisition and overall image stitching. First, a visual scanning sensor is used to measure various local areas of the topographic feature. Then, stitching technology is used to combine the different parts of the topographic feature to obtain a complete image.
This sensor's visual scanning probe is designed using the principle of local binocular stereo vision measurement. Overall topographic stitching essentially involves placing the collected data onto a common coordinate system, thus obtaining a holistic data description. Data is collected from above the measurement space at different angles and positions using a high-resolution digital camera. The spatial coordinates of control points are obtained using the principle of beam-oriented intersection adjustment, and a global coordinate system is established. Finally, the data is stitched together by associating and transforming the various coordinate systems.
V. Summary
Vision originates from a way in which organisms acquire information about their external environment. It is the most effective means for organisms in nature to obtain information and is one of the core components of biological intelligence. Humans acquire 80% of their information through vision. Inspired by this, researchers began to equip machines with "eyes," enabling them to acquire information by "seeing," just like humans. This gave rise to a new discipline—computer vision. By studying biological visual systems, researchers have attempted to imitate and create machine vision systems. Although these systems differ significantly from human vision, this represents a breakthrough in sensor technology. The essence of visual sensor technology is image processing technology. It captures signals from the surface of an object and creates an image, presenting it to researchers. The emergence of visual sensing technology has solved the problem of other sensors being limited by space constraints or the bulkiness of the detection equipment, making it widely welcomed by the industrial manufacturing sector.
