The 3D camera
3D vision in industrial image processing
3D (for the term “three-dimensional” or “three dimensions”) refers to the perception of depth in the space around us. To see in 3D, we use both eyes.
Humans have stereopsis (or stereoscopic vision or 3D vision), which means that they can perceive terrain and depth. Thus, each side of the brain receives information from both eyes: this is called bilateral visual projection. However, due to the distance between the two eyes and their apparent tilt, the two retinas perceive images slightly differently in two dimensions. It is this difference that allows the visual center of the brain to analyze, compare and generate three-dimensional images.
To understand how stereopsis works, you can fixate on an object by closing one eye then opening the other eye while closing the previously open eye. Note that each eye sees a slightly different image. When you open both eyes at the same time, your brain combines the two images to create the perception of a 3D image.
But how do you film a scene and represent it in three dimensions? To do this, it is necessary to have a camera capable of capturing the scene from two different angles. Some cameras have two lenses, sensors also called "dual cameras". Thanks to this type of camera, the two images reach the brain which combines them to create the perception of a 3D image.
The two lenses are used alternately, as if changing lenses on an SLR camera. By combining the two images produced we will be able to obtain better quality images, with more details, improved contrasts and less digital noise in low light situations.
(Panasonic 3D camera)
(Principle of stereopsis according to the Swiss medical journal - New developments in general internal medicine 2013)
The development of 3D cameras and 3D image processing technology opens up a wide range of application fields. In addition, consumer applications and software allow you to place 3D objects directly on a photo. The ease of use and the variety of 3D object libraries make them original and fun applications.
These cameras can be adapted to various applications and are used worldwide for industry and research.
3D mapping uses vision to profile three-dimensional objects and reliefs to translate them into the real world on a map. It is therefore among the latest methods of visualization and acquisition of information.
(3D mapping using a drone)
3D vision now also meets a wide range of needs for medical diagnosis. 3D models for MRIs, scanners or ultrasounds make it possible to precisely measure volumes and navigate inside the human body for observation and diagnosis.
In addition, certain cameras equipped with a 3D measurement system offer better control of quality control in the industrial environment, on the inspection of products based on criteria of shape, dimensions, detection of manufacturing defects, etc. The 3D vision evaluation allows increased precision for cases treated with difficulty by more traditional cameras.
Among the many technologies that have helped advance the robotics industry in recent years, 3D vision is one of the most impactful. Indeed, new imaging technologies, notably 3D cameras, are on the verge of unlocking capabilities previously inaccessible to robots.
Advances in robotics have enabled the creation of unique automated systems that not only replace humans in performing dangerous tasks, but also save lives. The main goal of their development is to improve the safety of rescue teams and expand their technical capabilities. Natural disasters and accidents pose a major threat to human life and property. In this case, whatever the danger, the survivors must be rescued as quickly as possible and provided with water, food and medicine. Rescue robots (eg: Robot Rescue) perform this complex task which would otherwise be very dangerous for the rescue team. Depending on the application, rescue robots can have different functions, structures or dimensions. 3D cameras check the quality of products, identify objects or detect obstacles without taking a scan unlike the 2D camera. In addition, the 3D camera offers very good image quality.
(3D modeling of a “pick and place” robot)
Robotization is a major challenge for production environments. However, many applications require the ability to guide the path of robots to account for dimensional variations from part to part. From this point of view, 3D vision is a very powerful tool, and recommended in particular for so-called delicate operations.
Today, 3D is everywhere: in cinemas, on television, on video game consoles, on cameras and even on robots. Many advantages seem to highlight the 3D camera such as the quality of the relief rendering or the precision provided for the analysis of a situation. But it must be kept in mind that the cost of such systems is higher than for more traditional cameras, particularly due to the greater number of components.
Apart from the 3D camera, other types of cameras have particular characteristics which can prove useful for use cases which the classic 2D camera could not respond to. In a future article, we will talk in particular about thermal cameras. How do these special cameras work? In what cases are they used? To find out, don’t hesitate to follow our publications.
Also consult
