Products & Services

RoboVision®2s Functions​ ​

Main functions of stereo vision RoboVision®2s

ZMP designs and manufactures everything from hardware to software. By using the attached application function, the distance from a photograph (image) taken by the camera to a specific object can be measured in a short time (high frame rate).
In order to recognize the distance in three-dimensional (3D) based on the parallax information by the image processing of the stereo camera, it is possible to detect and track objects in space with image processing software in real time and with high accuracy.

The following describes the functions of the stereo camera unit RoboVision®2s for hardware, software, and image processing applications.


RoboVision®2s is designed in-house. Upgrading from RoboVision®2 to RoboVision®2s, the optics parts were improved, the lens performance was improved, and an optical low-pass filter was added.
Improvement of the lens realizes improvement of the peripheral resolution, and the influence of moiré and false color is reduced by adding an optical low pass filter.
By using this product, the variation in the change of the measured distance value when measuring the distance has been reduced, and smooth distance measurement results can be obtained.
Improved RoboVision®2s lens



Developed a unique calibration method, convert the image measured by the calibration board into coordinates, convert it into an image viewed from the front, and extract edges to improve the edge detection accuracy of the checkerboard. We confirm the distance measurement result of 20m ± 1% accuracy before shipping.
Left: Shooting checker board Right: Recognition result

Image processing application

Measurement image and point cloud data output

The attached application outputs a simple image measurement, distance measurement result, and point cloud data (.pcd file). After purchase, you can connect a camera to a PC and easily perform development and research using image data and distance measurement results.
Diagram measurement image
Distance measurement result
Point Cloud Data

Object detection by point group (matching data)

Acquires three-dimensional point cloud information from parallax and distance measurement data, and outputs the presence or absence of an object, its position, and its width and height (size).
In addition, it detects objects on the road in front (plane) and outputs the results on a 2D map that looks down from above on information such as position, height, and width for each block (object / obstacle).
Our original ROGM (Redundant Occupancy Grid Map) improved from Occupancy Grid Map (OGM) enables stable detection of overlapping pedestrians as shown below.
* This software is sold as optional.

Figure Object Detection Application GUI
Figure RoboVision 2s CarTrack package GUI

RoboVision®2s Product Description Menu


For inquiries