Distance metering using laser
There are two main tecniques to measure the distance with laser beam. One way is to pulse the laser beam and measure the time between emitted and reflected pulse. Second way is to beam laser constantly and then shoot the camera with special filter to measure how much the dot of the laser has moved on obstacle. The latter is the cheaper method as it does not need that special processor that is capable of sampling at nanoscale timeframe.
Laser sensor characteristics
Model | URG-04LX-UG01 (Pulse measurement) | Parallax Laser Range Finder (CMOS measurement) |
---|---|---|
Typical accuracy | < 3 % | < 5% |
Input span | 0,02...4m | 15-122cm |
Power requirements | 500mA@5VDC | 150mA@5VDC |
Output form | digital | digital |
Weight | 160g | 10g |
Interface electronic circuits
Input signal handling depends on the technology how the measurement is implemented.
CMOS measurement
Figure 1: Distance measurement with angle detection method
If laser beam is projected inside camera focal area then it is possible to resolve the distance using simple math. The optical axis of camera and laser beam should ideally be aligned perfectly. However, it is possible to add constant angle correction into math if angle is not perfectly aligned. A relatively simple algoritm can be used to find brightest pixels. This method assumes that the brightest pixels are laser dot which is the case usually indoors but not necessarely outdoors. Thats why red pass filter should be used to lessen disturbance of other light sources. When laser dot is resolved from image we can calcute the distance because the dot position in the image frame is known.
Distance D can be calculated with formula:
In this formula the h is constant distance between laser and camera. Theta is calulated from:
So the final equation would be:
To cut out the amount of data that needs to be processed the image frame can be cropped to save only pixel area where the laser dot can be.
PSD measurement
PSD method is similar to CMOS but PSD is more accurate only matte or specular surfaces. While with CMOS one can choose the brightest pixel to be the center of the laser beam but with PSD the beam centroid is determined by whole reflected dot. If reflective surface is rough, the beam spreads quite large area and distance is calculated inaccurately. So PSD method suites well for example surveying where one need to use some specular stick anyway at the other end. However, that method is not so good for universal laser meter which is usually used with various materials which are often quite rough.
Figure 2 illustrates how different methods differ from each other. Of course this is the worst case scenario what can happen in certain conditions.
Figure 2: Accuracy comparison between PSD and CMOS
Both CMOS and PSD method can be used with highly reflective surfaces by sending laser beam at a angle that is right with the collector. There are devices available where the mode is user selectable. Without that mode the laser beam will reflect straight back where it came from but on the other hand if device is designed to be used only with specular materials the beam will not reflect to collector from matte surface.
Pulse measurement
Figure 3: Basic principle of time-of-flight (TOF) measurement
Noise
The amount of noise is dependant of meters bandwidth. The bandwith is defined usually by the point where output voltage is dropped 3 dB or 30%. The noise is distributed over all frequencies but it can be reduced by filtering which limits bandwidth.
References
https://sites.google.com/site/todddanko/home/webcam_laser_ranger
http://www.intorobotics.com/high-precision-laser-sensors-in-robotics/
http://www.parallax.com/sites/default/files/downloads/28044-LaserRangeFinder-v1.0.pdf
http://www.minifaros.eu/files/PaperAC11_KostamovaaraJ_OuluUniversity.pdf
http://www.mtiinstruments.com/technology/triangulation.aspx