Introduction to 3D Machine Vision


What is 3D Machine Vision?

3D Machine Vision is one of the most exhilarating technological advancements in the industry. It typically consists of multiple cameras with one or more laser displacement sensors. This system allows a robot to sense differences in its physical environment and adapt accordingly.

There will be an overall increase in flexibility, utility and speed in essential applications for factory engineers so they can accurately, quickly and cost-effectively sort out a large range of inspection challenges in an automatic production line.  The system not only sees its environment and orientation, but also calculates where the object is located at. These enables additional accuracy for depth perception which functions similarly to human eyes.

Main Forms of Technology to Generate 3D Images of Object

These system cameras have been used for detection of dirt, absence, presence and defects during the production process.

      1. Stereo Vision
Stereo Vision captures stationary objects. It utilises two cameras placed side-by-side that calibrates and focuses on the same object to provide a full 3D view of measurements in a dynamic and unstructured environment.

      2. Laser Triangulation (3D Profiling)
Laser Triangulation is a machine vision technique used to capture 3D measurements by pairing up with laser illumination source with a camera. It is based off triangulation rays of multiple perspectives using a camera that is perpendicular to the beam. Laser triangulation can provide a spectacularly detailed point cloud map of the object.

      3. Time of Flight (ToF)
Time of Flight refers to a range of imaging camera systems that employs time-of-flight techniques to resolve the distance between each point of the image and the camera. It measures the round-trip time of an artificial light signal originating from an LED or laser, which generates a point cloud based on the recorded times.

      4. Structured Light
Structured Light refers to the process of projecting light with a known shading pattern onto the captured scene. Its main purpose is to detect and measure deformities when scanning surfaces using the expected pattern on the scene, allowing vision systems to calculate the depth and surface information the objects have.

Benefits of 3D Machine Vision
Some benefits of the 3D machine vision technology is being able to capture more detailed information regarding an object, which is inclusive of its size, shape and position, essentially used to handle and carry out industrial operations. It is also among the best that produces automated image production and analysis. The 3D mapping outcomes of unforeseen variables, light or colour issues and positioning of the targeted objects are optimised as well, creating a more favourable circumstance for those using its systems.

It is labour-saving, increases efficiency, enhances productivity, reduces waste and increases compliances. With many industrial processes having elements of dull repetition, the system can relieve human labour from doing them, allowing the employees to focus more on other aspects that could benefit their organisation instead. It can also work for many hours compared to human labour due to fatigue issues.

The 3D machine vision systems made for industrialisation purposes also increases their productivity and efficiencies through cost reductions and improvements in the quality of products. The accuracy of the continuous inspection increases with the use of systems, enabling quicker and more precise identification of undamaged parts and pieces. This will boost the overall reduction in wastes.

These systems also allow easier methods of reviewing processes and tracing back to problems that are important for compliance. If a defect is found on a product, one will be able to review the process by referring to images captured during the production phase to narrow down the issue and make corrective decisions from the source.

2D Machine Vision VS 3D Machine Vision
2D machine vision was first discovered in the 1950s in its early form and uses a digital camera to capture images of an object. It processes based on comparing variations in intensity of contrast. It provides area scans that works well for discrete parts. Most software packages are compatible with these systems and are widely used in machine vision applications. The initials of 3D machine vision was developed in 1978 and in recent years, evolved into a very powerful technological tool commonly found in manufacturing environments.

3D Machine Vision is a mighty system that enables more accuracy for localisation, recognition, and inspection tasks than traditional 2D Machine Vision. It is able to process issues in greater depth, providing solutions that 2D systems cannot resolve. 2D cameras usually takes an image of light reflected from the object. This is why when changes are made in the illumination, it can cause adverse effects on accuracy when taking measurements.

2D machine vision systems have limitations in environments where illumination cannot be controlled or altered to fix a shot. Too much light causes overexposed shots while too little affects the clarity of edges and features appearing on 2D imaging. Also, even though 2D machine vision systems are great tools in many applications and have been used for decades, it has a few fundamental problems due to only being able to see flat images.

Unlike 2D machine visions, 3D machine visions can inspect the subject as they are and measure them up to scale, without the extra need of applying complex algorithms and heuristics, which assumptions often fail, leading to errors in applications in 2D machine visions. In the context of applications involving dimensioning, thickness measurements, space management and depth quality control, 3D machine vision outperforms 2D machine vision by a landslide.