With the advent of the Industry 4.0 revolution, the demand for 3D vision is increasing due to the limitations of 2D vision in complex object recognition, the accuracy of dimension annotation and distance measurement, as well as its limitations in complex situations such as human/robot interaction.
3D vision enhances the autonomy and effectiveness of robot/machine systems in the factory automation market, as it is crucial for applications limited by 2D vision, such as higher-precision quality inspection and reverse engineering. Furthermore, the use of vision-guided robots is growing, requiring 3D vision for better remote guidance, obstacle recognition, and precise movement.
3D vision also uses systems that can prevent and resolve dangerous situations, as well as surveillance systems that can count and distinguish between people and robots or objects, to protect the safety of factory workers in intensive human-machine interactions.
3D vision is impacting society by enabling safer, better-performing, and more efficient driver assistance systems for end users. For example, 3D vision is a driving force behind advanced driver assistance systems (ADAS) such as autonomous vehicles and collaborative robots.
2D imaging remains prevalent in fields such as barcode scanning and OCR. It plays an indispensable role in factories and warehouses, where its use is increasing thanks to the adoption of blockchain and the booming e-commerce sector driving phenomenal growth in logistics centers and transportation. Teledyne e2v offers unique 2D imaging products specifically designed for code scanning, such as the Topaz sensor series, characterized by high scan rates and reliability.
There are various techniques and methods for obtaining 3D images. The main ones are:
● Stereoscopic vision – This method uses two cameras mounted at different angles to an object, employing calibration techniques to align pixel information between the cameras and extract depth information. This is similar to how our brains measure visual distance.
● Structured light - A known light pattern is projected onto an object, and depth information is calculated by how the pattern deforms on the object.
● Laser Triangulation - Laser triangulation systems achieve 3D measurement by pairing a camera with a laser source. Since the angular offset between the laser and the camera is known, the system uses trigonometric functions to measure the geometric offset of the laser line (its value is related to the height of the object). This technique is based on scanning the object.
● Time of Flight - The light source is synchronized with the image sensor, and the distance is calculated based on the time between the light pulse and the light reflected back to the sensor.
Each technology has its advantages and disadvantages; therefore, depending on the target application (especially distance range and depth accuracy requirements), some technologies are more suitable than others. A relative comparison is provided in Table 1.
Table 1: Top-level comparison of 3D imaging technologies
While it still represents a small segment of factory automation and warehouse vision systems, a growing number of 3D systems are now based on 3D stereo vision, structured light cameras, or laser displacement. These systems operate at fixed working distances and require extensive calibration of specific inspection areas. Time-of-flight systems overcome these challenges, offering greater flexibility from an application perspective, but most are still limited by image resolution.
Teledyne e2v has a strong foundation in machine vision, including inline scanning cameras and area scanning sensors, and has now established a unique platform specifically for 3D imaging. This will support the latest industrial applications such as vision-guided robots, automated guided vehicles (AGVs) for logistics automation, factory monitoring and security, handheld scanners, and outdoor applications. Teledyne e2v aims to provide consistent products based on a variety of 3D technologies to meet customers' application needs.
