High-Precision Surface Inspection: Uncertainty Evaluation within an Accuracy Range of 15μm with Triangulation-based Laser Line Scanners

Abstract

Triangulation-based range sensors, e.g. laser line scanners, are used for high-precision geometrical acquisition of free-form surfaces, for reverse engineering tasks or quality management. In contrast to classical tactile measuring devices, these scanners generate a great amount of 3D-points in a short period of time and enable the inspection of soft materials. However, for accurate measurements, a number of aspects have to be considered to minimize measurement uncertainties. This study outlines possible sources of uncertainties during the measurement process regarding the scanner warm-up, the impact of laser power and exposure time as well as scanner's reaction to areas of discontinuity, e.g. edges. All experiments were performed using a xed scanner position to avoid e ects resulting from imaging geometry. The results show a signi cant dependence of measurement accuracy on the correct adaption of exposure time as a function of surface re ectivity and laser power. Additionally, it is illustrated that surface structure as well as edges can cause signi cant systematic uncertainties.

Measuring Leaf Thickness with 3D Close-Up Laser Scanners: Possible or Not?

“…This sensor works according to the light-section method and principally measures the distance between the sensor's reference and the object's surface in two dimensions [17,18]. Therefore, a red laser line (wavelength: 660 nm) is emitted, backscattered from the object's surface and imaged by two CCD (charge-coupled device)-arrays.…”

“…However, uncertainties of the LTS mainly depend on an accurate adaption of the sensor properties, e.g., the exposure time to the optical characteristics of the object's surface [18]. In order to reach the highest precision of the distance measurement, the exposure time has to be adjusted in a way to guarantee an optimal signal-to-noise ratio.…”

“…From a metrological viewpoint, this indicates a displacement of the laser line and, consequently, a distance measurement that is systematically too long. Furthermore, the intensity of the signal is weakened by the chlorophyll, which in turn leads to a higher measuring noise [18]. A further source of uncertainties can be attributed to the angle of incidence [24,25].…”

“…Both point clouds were approximated by the best fitting plane using an established least-squares adjustment approach [18,27]. The mean distance between the two measured surfaces, i.e., the thickness, is calculated from the distance of each point of side A to the plane of side B and vice versa.…”

“…Thus, the exposure time of the LTS was adjusted automatically to the reflective properties of the gauge's surface by applying the sensor integrated procedure. During the measurement, the scanner-surface distance and the angle of incidence were kept constant and the laser line was oriented orthogonally to the edges to minimize systematic uncertainties [18]. The results are two point clouds of the upper and bottom side of the gauges (Figure 4).…”