Author: Zhu Jincan
Source: clever101's column
As the name suggests, visual SLAM (aka vSLAM) uses images collected from cameras and other image sensors. Visual SLAM can use ordinary cameras (wide-angle, fisheye, and spherical), compound-eye cameras (stereo and multi-camera), and depth cameras.
normal camera
There are three main types of common cameras used in visual SLAM: wide-angle, fisheye, and spherical cameras.
Wide-angle camera, as the name suggests, refers to a type of camera with a lens with a wider viewing angle, that is, when a wide-angle camera shoots an image, it can capture more scenes within a limited distance. Among them, the size of the minimum focal length is an important basis for determining whether the camera is a wide-angle camera. Generally, cameras with a minimum focal length of less than 28mm are wide-angle cameras. Generally speaking, a wide-angle camera is actually a camera with a wide-angle lens. It is a camera with a short focal length, a wide viewing angle and a deep depth of field. It is a type of camera that appears specifically for taking photos of large scenes.
A fisheye camera refers to a camera with a fisheye lens, which is a lens with an extremely short focal length and a viewing angle close to or equal to 180°. A lens with a focal length of 16mm or shorter. It is an extreme wide-angle lens, "fisheye lens" is its common name.
Spherical camera, also known as dome camera, is mostly used in security monitoring occasions, and its appearance is as follows:
compound eye camera
Compound eye cameras mainly include stereo cameras and multi-cameras. Stereo camera refers to a camera with two or more image sensors, which enables the camera to simulate human binocular vision, thereby enabling it to perceive depth. Simply put, a stereo camera can generate at least two left and right images to form a stereo pair.
Multi-cameras, also known as multi-camera arrays (Camera Arrays), use multiple cameras at different spatial positions to collect photos from different perspectives. A stereo camera is a special case of multiple cameras.
depth camera
Depth cameras, as the name suggests, can measure the distance (depth) of an object to the camera. Traditional RGB color ordinary cameras are called 2D cameras, which can only capture objects within the camera's viewing angle, without the distance information from the objects to the camera, and can only perceive the distance of objects by feeling without clear data.
RGB-D depth camera (also known as 3D camera, where D stands for Depth is the depth information) can obtain the distance information from the object to the camera, together with the X, Y coordinates of the 2D plane, the three-dimensional coordinates of each point can be calculated, so we Applications of depth cameras such as 3D reconstruction, object localization, and recognition can be inferred. At present, there are three mainstream depth cameras: structured light, time-of-flight, and binocular stereo.
references
1. What is a stereo vision camera?
2. Multi-camera array overview
3. Depth camera: structured light, TOF, binocular camera