High Accuracy Stereo Vision


3D camera model for Unlimited field of view

No Need for Computation of Projection Matrix

From ultra wide to narrow field of view. Stereo vision also supports different types of cameras to be used simultaneously .

Two Points Calibration

Only two points are required for a stereo vision system calibration.

Any Geometry

Mutual distance and location between cameras is not limited. Up, down, left, right, in front, behind.

High Accuracy

Simple two point calibration procedure enables 3visionD stereo vision system more than 99% accuracy,

Novel approach to Stereo vision

Video Camera 3D Model

3visionD has developed unique method that uses camera parameters to create camera 3D model. This model is then integrated into real 3D world. 

In Smart Parking application video cameras are integrated with open street maps (via geo coordinates) and vehicles detected in video stream are geo tagged. In this case 3D camera model is applied on existing cameras. 

Stage tracking application is using ultra wide angle camera with 150 degrees field of view to track the stage performer with moving lights. In this case, 3D camera modes is using our lens distortion algorithm. 

3visionD Plugin for Milestone XProtect Video Management Software uses geo coordinates, Cartesian coordinates and lens distortion models for new and existing cameras. It detect people, vehicles, pets for geo enabled alarming. 

Video camera 3D model has prove its accuracy and reliability in all these applications so the next logical step is stereo vision. 


What is the difference ? 

Video Camera 3D Model

3visionD stereo vision produces the same output as standard stereo vision system. The x,y,z coordinate of targeted point.

What is the difference?

1. 3visionD does NOT use reprojection matrix. System is calibrated by using two calibration points and three distance measurements. 

2. Therefore, chessboard pattern calibration is not required.

3. Any camera cam be used for stereo vision. When ultra wide angle cameras are used, lens distortion file is used so that system does not require chessboard calibration on site. 

4. Any combination of cameras can be used. 

5. 3visionD stereo vision uses system geometry. Distance between cameras or 3D camera mounting geometry must be known. 

6. Two point system calibration. Only distances between two points are measured for system calibration on site. 


Video camera as 3D Model

Video camera is defined as a 3D shape. 3visionD proprietary technology takes camera, lens and image parameters creates and unique 3D shape that describes video camera.

Camera Matrix

Camera and lens parameters are stored as metadata files. This files holds all the data required for camera calibration, lens distortion and etc. Therefore chessboard pattern calibration is not required.

Any camera as Stereo camera

3D video camera model has defined video camera as unique geometrical shape and this shape is enabling the combination of different camera types to be used in one stereo vision system.

Basic Two Point system calibration

Camera misalignment is compensated by two point calibration. Misalignment between center of cameras field of view is measured in mm for automatic compensation.

Any geometry

3visionD stereo vision supports any geometry between cameras.

System setup by technician

System setup does not require skilled profesional.


3visionD SIA

Brivibas gatve 300-9, Riga LV-1006, Latvia

3visionD is Supported By