Measuring Principle

Laser Triangulation

The device acquires height profiles and height images based on the laser triangulation principle. According to this method a laser line is projected on the target object from one direction. The image sensor views the object from a different but known angle. The resulting sensor image is evaluated by the embedded processor and converted into a single height profile. By moving the target under the laser line with a certain speed, a complete height image can be acquired. The three points, laser, sensor and target and the angles and distances between them define the triangulation geometry.

Geometric Dependencies

There are three major geometries on that the whole device is build and calibrated:

Field of View

The Field of view (FOV) is dependent on the built-in optics of the device and the laser fan angle of the laser. The Near FOV is defined at the upper limit of the Z-range and is narrower than the FOV. The Far FOV is defined at the lower limit of the Z-range and is wider than the FOV.

Note

The laser line is usually wider than the FOV. This is necessary due to inaccuracies at the edges of the laser line.

Working Distance

The working distance (WD) describes the nominal distance between the bottom edge of the laser module and the target surface that needs to be measured. The accuracy of the measurement will always be the best when it’s performed at the working distance. However, there is a Z-range defined in which the target surface or multiple surfaces can appear. The total Z-range is the sum of the distances from the Near FOV to the WD and the WD to the Far FOV.

Triangulation Angle

The triangulation angle describes the angle between the vertically aligned laser and the tilted sensor plane.

Coordinate Systems

The coordinate system of the device is described as followed: The X-axis describes the width of the measured area along the laser line, the Y-axis describes the transport direction and the Z-axis describes the height values along the laser plane.

The figure below demonstrates the typical triangulation geometry.

Resolution

The resolution of the sensor is different in each axis:

ΔX : Resolution along the laser line and across the target (lateral). It is the FOV width divided by the number of pixels of the imager
ΔY : Resolution perpendicular to the laser line (longitudinal in the direction of motion). It is directly dependent on the measurement frequency and the transportation speed.
ΔZ : Height resolution. The laser line is projected perpendicular to the object surface, while the camera views the object under the triangulation angle α. The height resolution can be approximated by:

Occlusions

Occlusion is a major limitation of the laser triangulation method. If the laser line is not visible by the sensor, no height data can be acquired. It is necessary to analyze the target object and to plan the scan path to avoid occlusions. Using a dual-head or multi sensor configuration can be a solution.

Device modes and peak detection algorithms

This chapter describes the general functionality of the implemented device modes and peak detection algorithms. This device can be operated in a 2D Areascan mode and a 3D Linescan3D mode:

Areascan

The Areascan mode is the 2D image mode in which the device is operated similar to a standard 2D camera. In this mode grey scale data of 8- to 10-bit resolution is acquired over the device interface. Furthermore, the sensor can be divided into multiple regions.

Linescan3D

The Linescan3D mode is the 3D mode, which delivers height data. Just like in the Areascan mode the sensor can be divided into multiple regions. For each region it is possible to enable a Scan3dExtraction. The Linescan3D mode can be set to different peak detection algorithms. In this mode the image sensor takes one image from which the selected algorithm extracts the laser line position column wise. All laser line positions from one image are called a profile. It represents a cross section of the target object. The device repeats this process for a defined number of profiles which are then combined into one range image. This range image does not represent an image of the target object in a classical sense but rather a set of height profiles.

3D Algorithms