2.5D Machine Vision System

Share Info & Ideas
, , ,
1

A clever extension for B&R smart sensors and cameras uses laser triangulation with FPGA preprocessing to deliver a cost-optimized solution for solving 3D tasks in mid-range applications. B&R’s 2.5D machine vision solution captures moving objects to enable height monitoring, sorting, volume estimation and more. Machine builders can simplify system complexity with an integrated system – e.g. no need for a transducer. Vision can be synchronized to highly dynamic movements with nanosecond precision.

image
image993×460 77.7 KB

What is the functionality of the system in principle?

Laser triangulation is a method used to measure distances using a laser beam. A simplified explanation how laser triangulation works:

image
image517×588 71.9 KB

How can you imagine it?

image
image1089×343 38.2 KB

The initial image illustrates a scenario in which various boxes of differing heights are being transported via a conveyor belt.
The second image shows the same image with different grey values.Each grey value represents a height information, and the Vision Function 2.5D Blob allows you to analyse the image in terms of 2D information and height information.
The third image shows a different type of representation, which clearly illustrates how the high information is a gray value image.

Processing by using VF 2.5D BLOB.

High variability in inputs and outputs of the Vision Function 2.5D BLOB for the configuration of the system:

image
image1048×1305 241 KB

Use Cases

  • Volume measurement
  • Height measurement
  • Component position detection with low contrast between the component and the background

Example

image
image1600×836 65.1 KB

HW/SW overview

Variants

  • 3 mechanical sizes (length of the carrier rail = 450, 550 and 800 mm)
  • 3 laser variants (60°, 75°, 90°)
  • 3 different camera angles (adjustable on each carrier rail)
  • 2 different scan areas

This results in 18 measuring ranges:

  • Width area: 260mm up to 1.3m → Resolution 210µm to 1.02mm
  • Height area: 55mm up to 828mm → Resolution: 240µm up to 5.6mm

HW:

  • Smart Sensor/Camera
    • VSC122xx2.04xx-000
    • VSS122xx2.04xxP-000
  • Mechanics
    • VMA025D30
    • VMA025D40
    • VMA025D65
  • Lasers
    • VLA004075.660
    • VLA001060.660
    • VLA007090.660

SW:

  • since mapp Vision 5.28

License overview

  • mapp Vision “Premium”
  • mapp Vision “Basic”

QnA

  • Why is it called 2.5D and not 3D?
    • we provide the height information, in a gray value image.
    • we don’t support a complex 3D point cloud.
    • Note, in the field of machine vision, the term 3D vision is usually associated with a stereo vision camera.
      A stereo vision camera is used for applications such as bin picking applications and can’t be compared with laser triangulation technology - like our 2.5D machine vision system.

:books: Related topics/additional info

:link: Automation Help

8 Likes
2

Thanks @kamil.ruzicka for the description of this innovation developed by the B&R vision team. Another interesting topic published in the past was anomaly detection. Check it out! Anomaly Detection for Machine Vision

4

Is this solution only suitable for measuring one object at a time, or is it possible to find multiple objects that are in the same image? In our case, we often have a tray with several similar objects that are to be picked with a robot and currently we use an external vision system to find the coordinates of these objects. Would it be possible to use this solution instead?

5

The solution can measure multiple objects at a time. However you have to keep in mind the possibility that the objects could shade each other and effect the results.

I would suggest contacting your B&R representative for more detailed information regarding your use case.

1 Like