小优智能科技有限公司成立于2015年底,是一家专注于高精度3D机器视觉模组研发、生产及销售的高科技企业。
公司自主研发的3D机器视觉模组采用激光/DLP白光编码光栅结构光+双工业相机方案,还原物体三维信息,广泛应用于消费电子领域、工业领域和安防领域,具有精度高、速度快、成本低的优势。
At present, the depth camera on the market mainly includes: binocular camera, TOF, structured light scheme.
Binocular camera:
Principle: Based on the parallax principle and the use of imaging equipment to obtain two images of the measured object from different positions, by calculating the position deviation between the corresponding points of the image, to obtain the three-dimensional geometric information of the object.
Advantages: cheap; High resolution; High frame rate;
Disadvantages: High requirements for stereo matching algorithm, which requires a large amount of computation; Not suitable for scenes with little texture;
TOF:
Principle: The TOF system is actually a liDAR system, which can emit light pulses from the transmitter to the object, and the receiver can determine the distance of the measured object by calculating the travel time of the light pulse from the transmitter to the object and then back to the receiver in pixel format.
Advantages: Good measurement accuracy and stability; Little affected by the environment, good anti-interference performance; Wider viewing Angle; Higher real-time
Disadvantages: High cost, low resolution
Structured light Scheme:
Principle: The light with certain structural characteristics is projected onto the object to be photographed, and then collected by a special infrared camera;
Advantages: small resource consumption, low cost, high precision in a certain range.
Disadvantages: easy to be affected by ambient light; The distance measurement effect is poor.
1. Binocular matching 3D imaging principle
Binocular is mainly dependent on matching algorithm
2. Structured light imaging
Point structured light
For this corresponding parameter, we first have to solve for the distance between two angles and a baseline.
Linear structured light is relatively familiar, so I won't say;
Principle of optical coding imaging
I've heard this before, but I don't know much about it.
Principle: Project a surface onto an object, receive the camera through the pattern encoding to identify which Angle of the projector "ray", and calculate the distance. It feels a little bit like the point of view.
TOF
principle