study of ambient light influence on laser 3d scanning

327

1. Introduction

3D scanning is an emerging technology, used mostly in new product development. Number of vendors offering more affordable 3D scanners is constantly increasing. Both hardware and software solutions compete with each other on the global market, increasing accuracy, speed and resolution, decreasing simultaneously related costs and becoming more user-friendly. One of the popular low-end 3D scanners is shown in Figure 1. It has measurement declared measurement uncertainty of 0,12 mm, and it is shipped with rather simple scanning software.

Fig. 1. Desktop laser 3D scanner NextEngine.

Fig. 2. Scanning software for NextEngine 3D scanner.

One of the first problems which novel user faces using this software is choice of adjustable parameters in order to obtain the desired scan with as little steps as possible. The user has an opportunity to choose between “matte” and “shiny” surface, and colour between “dark” and “light”. This choice is not always too intuitive and has a strong influence on scanning result, especially in terms of time needed. For example, it is not always obvious whether red is “darker” than green, or blue is “lighter” than gray.

This research was conducted in order to speed-up the procedure of setting these parameters. Our intention is to give recommendations on how to set these parameters for some typical scanned objects.

STUDY OF AMBIENT LIGHT INFLUENCE ON LASER 3D SCANNING

S. Lemeš, N. Zaimović-Uzunović

University of Zenica, Faculty of Mechanical Engineering, Fakultetska 1, 72000 Zenica, Bosnia and Herzegovina

AbstractThe quality of 3D scanned data is influenced by many factors, such as scanned surface colour, glossiness, geometry, ambient

light, scanner resolution, proper selection of scanned segments, etc. Ambient light is important source of errors, especially for scanners that are based on projected patterns and structured lighting. The laser 3D scanners are also influenced by ambient light, and this paper studies the effects of ambient light on a specific commercial scanner. Preparatory phase of digitising process includes selecting proper combination of scanner’s sensor aperture and shutter time, depending on the characteristics of scanned surface. This paper investigates a relationship between the ambient light intensity, the colour of scanned surface, and the laser scanning quality.

Keywords: 3D scanning, Ambient light, Laser



328

2. Related works

Voisin et al. in [1] studied ambient light influence for 3D scanners based on structured light. They showed that ambient light can introduce errors into scanned data from most commercial 3D scanners based on projected patterns and structured lighting. They also proposed a physical explanation for the systematic error observed from coloured patches.

Voegtle et al. in [2] studied influence of different realistic object materials and object colours on the measurements of a terrestrial laser scanners. They showed that grey scaled test plates have proven a significant dependence between the brightness of the scanned object and the obtained accuracy. Another important result is the enormous difference between measurements at day and night-time (approx. a factor of 2).

Lichti and Harvey in [3] studied influence of reflecting surface material on time-of-flight laser scanner measurements. Their results showed no significant range errors due to differing material properties, but they did exhibit changes in range measurement distribution and return signal intensity.

Amiri Parian and Gruen in [4] performed an accuracy test of the point cloud of the laser scanner by using the laser’s intensity image. They focused on other sources of errors, such as eccentricity of the scan centre, collimation and horizontal axis error, tumbling error and resolution of horizontal and vertical rotation.

Boehler and Marbs in [5] investigated laser scanner accuracy through different test targets. They proposed the deviation of single points from the object’s surface as an indication for the accuracy. They also studied influences of surface reflectivity, environmental conditions and edge effects.

3. Experiment

3.1 Equipment used

For the purpose of this research the NextEngine desktop laser 3D scanner was used. The technical specifications of this scanner are listed in Table 1.Table 1. NextEngine 3D scanner specifications

Laser Class 1M, 10 mW, solid state, 650 nmSensor 3.0 Mpixel CMOS Resolution 400 DPI on target surfaceAccuracy ± 0.005” (± 0.13 mm)

The ambient light intensity was measured by analogue exposure meter Lunasix by Gossen, Germany (Fig.3).

Fig. 3. Exposure meter Lunasix.

3.2 Target surfaces

In order to have accurate colours of scanned target surfaces, standard samples of coloured surfaces were used. The samples were manufactured by MACtac Europe S.A., Belgium (Fig.4), and they have RAL standard colour codes. RAL is a colour matching system used in Europe, mainly for varnish and powder coating, but also for graphics and CAD software. Table 2. summarizes RAL colours of samples used in this experiment.

Fig. 4. Samples used as scanning target surfaces.

Table 2. RAL colours used, with RGB (0-255) codes [6]

RAL MACal colour name R G B9010 6622 Glossy White 240 238 2269005 6623 Glossy Black 188 19 201016 9809-6 Glossy Yellow 240 232 643020 9859-04 Glossy Red 190 17 166029 9849-10 Glossy Green 0 110 595017 9839-22 Glossy Blue 0 81 1407042 9889-03 Glossy Grey 141 145 1457012 9889-03 Glossy Grey 87 93 94



329

3.3 Data processing

The target surfaces with different colour and under two different light intensities, were scanned at the same position, and with different settings of colour (dark-light). The scanned point clouds were exported into OBJ data formats and trimmed to fit the same enclosed volume. To estimate the scanning quality, points were counted and compared to theoretically maximum number of points. Lower number of points referred to unsatisfactory colour/surface settings. Figure 5 shows examples of scanned surface with satisfactory and unsatisfactory number of points.

Fig. 5. Scanned samples

4. Discussion on results

Figures 6 to 12 show results of data analysis performed: they show how surface settings set in software (from 5:dark to 100:light) influence the number of points scanned. Each colour sample was scanned under normal daylight (120 lux), and in dark (20 lux). The dark conditions were accomplished by covering scanner and scanned object with large cartoon box.

Fig. 6. Scanning results for white sample (RAL 9010)

When white surfaces are scanned, ambient light decreases number of points, due to strong reflection.

Fig. 7. Scanning results for black sample (RAL 9005)

When black objects are scanned, ambient light has weak influence onto scanning result.

Fig. 8. Scanning results for yellow sample (RAL 1016)

Yellow surface shows almost identical results as the white one; i.e. reflected ambient light reduces number of usable points.

Fig. 9. Scanning results for red sample (RAL 3020)

Red surface acts interestingly, possibly due to red laser light used. When scanned in daylight, with surface settings set below 20, no points are scanned.



330

Simultaneously, when scanned in dark room, with surface settings set below 20, about 20% of points appear in scanned result.

Fig. 10. Scanning results for green sample (RAL 6029)

Green surface is rather sensitive to ambient light.

Fig. 11. Scanning results for blue sample (RAL 5017)

Ambient light has no influence on blue surface.

Fig. 12. Scanning results for light grey sample (RAL 7042)

Light grey surface shows absolutely best scanning results. No matter how surface settings are set in software, or is there ambient light or not, the scanner gives more than 93% of points in each scan.

Dark grey (RAL 7012) surface is also insensitive to ambient light. There are no differences between the results obtained under different lightning conditions, and surface settings in Nextengine software should be set between 5 and 70%, which gives almost 100% scanning results, similar to blue surface (Fig. 11).

5. Conclusions

The influence of ambient light, surface colour and materials on the 3D scanning was investigated by number of authors [2, 3, 4, 7]. These influences were tested on long-range or geosurvey scanners, and this is the first effort to test it on a desktop 3D laser scanner.

Contrary to common opinion, the ambient light does have influence onto laser-based 3D scanners. This influence is especially strong when glossy white, yellow and green surfaces are scanned.

To achieve the best scanning results, it is desirable to have light grey surface, since it is the least sensitive on ambient light and program settings.

This research should be expanded to include other influence factors, such as surface glossiness, different anti-reflection coatings, laser beam colour, etc.

References

[1] Voisin S., Foufou S., Truchetet F., Page D., Abidib M.: Study of ambient light influence for threedimensional scanners based on structured light, Optical Engineering 030502-1, March 2007/Vol. 46(3)

[2] Voegtle T., Schwab I., Landes T.: Influences Of Different Materials On The Measurements Of A Terrestrial Laser Scanner (TLS), The Int. Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. 37. Part B5. Beijing 2008

[3] Lichti D., Harvey B.: The effects of reflecting surface material properties on timeof-flight laser scanner measurements, Symposium on Geospatial Theory, Processing and Applications, Ottawa 2002

[4] Amiri Parian J., Gruen A.: Integrated Laser Scanner and Intensity Image Calibration and Accuracy Assessment. In: International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. 36, Part 3/W19. 2005

[5] Boehler W., Bordas V. M., Marbs A.: Investigating Laser Scanner Accuracy. In: Proceedings of XIXth CIPA WG 6, International Symposium, Antalya, Turkey, 30. September– 4.October, pp. 696 -702. 2004

[6] http://www.easyrgb.com/index.php?X=TINT[7] Clark J., Robson S.: Accuracy of measurements made

with a Cyrax 2500 Laser Scanner against Surfaces of known colour. Survey Review, 37 (294). pp. 626-638. ISSN 00396265, 2004

study of ambient light influence on laser 3d scanning

Documents