Typical Application Areas of 3D Cameras:
1) Detecting obstacles and performing "manual" navigation for autonomous vehicles (such as forklifts) in industrial environments
2) Robot-controlled conveyor belt grasping or parts picking operations
3) Performing presence/absence detection, even when objects in containers/boxes have virtually no contrast compared to the background, allowing for inspection and counting
4) Position and presence/absence inspection of components on circuit boards
5) Volume measurement of various objects, etc.
Popular 3D Technologies
Commonly used technologies in 3D image processing include:
1) Stereo vision and structured light
2) Laser triangulation
3) ToF (Time-of-Flight)
Stereo Vision and Structured Light
Stereo vision works on the principle of human binocular vision. Two cameras are used to record two 2D images of an object. By recording the same scene from two different positions, a three-dimensional image is synthesized using depth information based on the principle of triangulation.
Stereo vision uses image data from two standard 2D array cameras to provide depth values for a scene. The images are adjusted based on the camera positions and applied geometric information. After adjustment, a matching algorithm searches for corresponding points in the left and right images to create a depth map of the scene.
The working distance of this method depends on the baseline (the distance between the cameras), and therefore varies depending on the application.
One way to improve the performance of a stereo system is to add structured light to the stereo solution. By using a light source to project bright geometric patterns onto the scene, the accuracy of the measurements can be improved, significantly reducing stereo imaging defects caused by homogeneous surfaces and low light. By calibrating the projector and the camera, it is even possible to eliminate the need for a second camera.
Stereo vision can achieve high accuracy. Difficult surfaces do not significantly affect stereo vision, but it always requires the object to have a small number of reference markers or random patterns. This means that this technology is generally not well-suited for use in production environments. Typical applications of stereo vision include: coordinate measurement technology, 3D measurement of objects and workspaces in industrial, service, or robotic systems, and 3D display of hazardous or inaccessible workspaces. Stereo systems are also well-suited for use in outdoor measurement systems, such as measuring and inspecting tree trunks in sawmills.
However, if a high processing load, complex installation work, and higher costs are acceptable, stereo vision is also suitable for industrial applications of target measurement when structured light is added.
Laser Triangulation
Laser triangulation combines the use of a 2D camera and a laser light source. In this process, the laser projects a line or point onto the scene in front of the camera.
The laser line or point appears on the object in front of the camera and is recorded by the 2D camera. If the camera is moved across or alongside the target (e.g., via a conveyor belt), the distance between the measured object and the sensor changes, and the viewing angle of the laser line or point changes with its position in the camera image. Thus, through mathematical calculations, the distance between the object and the light source can be calculated from the position coordinates in the image.