Calibration of Galvanometric Laser Scanners Using Statistical Learning Methods

Lüdtke, Stefan; Wagner, Benjamin; Bruder, Ralf; Stüber, Patrick; Ernst, Floris; Schweikard, Achim; Wissel, Tobias

doi:10.1007/978-3-662-46224-9_80

Cited by 5 publications

(7 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is in line with [26], where it was shown that for stepwise regression only a degree of 4 or 5 was needed to achieve optimal results. Ridge regression allowed polynomials up to degree 9, but regularizes the weights of w for stable results in a very high dimensional feature space (w R 9000 ).…”

Section: B Polynomial Modelssupporting

confidence: 88%

“…Both require an additional triangulation step, while we show that data-driven approaches do not. We compare these two existing approaches with artificial neural networks (ANNs) and ridge regression [21], [26]. For the first time we also point out possible weaknesses of ANNs and indicate under which conditions a consideration of other data-driven approaches such as Support Vector regression (SVR) or Gaussian Processes (GPs) may be fruitful.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Data-Driven Learning for Calibrating Galvanometric Laser Scanners

et al. 2015

Self Cite

View full text Add to dashboard Cite

State-of-the-art calibration very often constructs models motivated by a real-world device. Recently, artificial neural networks (ANNs) have been proposed as a more universal, accurate, and practical black-box approach. For a galvanometric triangulation device based on two mirrors, we embrace this proposal and set it into context with other supervised data-driven approaches: 1) ridge regression; 2) support vector regression; and 3) Gaussian processes. We show that they outperform available model-based approaches and yield similar performance compared with a memorizing lookup table calibration. The results demonstrate that an off-the-shelf usage of ANNs may run into generalization problems. Restricting the space of functions using kernel-based learning has proven to be advantageous. Finally, all approaches and distinct properties are discussed in a broader context, since each application entails differently relevant requirements for its calibration. This also holds for any calibration other than the considered triangulation device.

show abstract

Section: B Polynomial Modelssupporting

confidence: 88%

Section: Introductionmentioning

confidence: 99%

Data-Driven Learning for Calibrating Galvanometric Laser Scanners

et al. 2015

Self Cite

View full text Add to dashboard Cite

show abstract

“…For instance, Mao et al [8] and Yang et al [9] reported the calibration of one-mirror galvanometric scanners, which were used in 3D measurement systems. Lüdtke et al [10] and Tu et al [11] calibrated two-mirror galvanometric laser scanners by using employed statistical learning methods (e.g., linear regression and artificial neural networks). For two-mirror galvanometric scanning systems, they were generally calibrated by calculating correction tables for all joint componentsi.e., two rotatable mirrors.…”

Section: Literature Reviewmentioning

confidence: 99%

In situ calibration of galvanometric scanning head for laser powder bed fusion machines based on a vision system

Quinsat

2020

Int J Adv Manuf Technol

View full text Add to dashboard Cite

In Laser Powder Bed Fusion (LPBF) additive manufacturing machines, a laser galvanometric scanning system is used to control the movement of the laser beam, which melts material powders layer-by-layer in a powder bed. Errors between desirable positions and actual positions of the laser in the work plane of the machine directly affect the geometry accuracy of built parts. Therefore, calibrating the laser galvanometric scanning system of LPBF machines is a very important task. This system is generally calibrated by calculating correction tables for its joint components. However, most existing calibration methods require a measuring phase of actual positions of laser engraved on a calibration plate. This phase was generally performed by a measuring system in another location, thus requiring a delay time and causing an interruption of the calibration process. In the current study, an in-situ calibration method for the laser galvanometric scanning head of LPBF machines was developed based on a vision system. The vision system (i.e., a high-resolution monochrome camera) was used to acquire experimental data rapidly during the calibration process, thereby avoiding the laser marking and the measuring phases of laser marks in another location. The acquired experimental data was subsequently used to calculate the correction tables/matrices for calibrating the optical chain of an LPBF machine prototype. The obtained results show that the developed measurement method is acceptable for the in-situ measurement with a good accuracy. After the calibration process, errors between the positions generated by the identified LPBF machine and those measured by the proposed method are very small with an average error of 0.03 mm, and about 80% of positions present an error less than 0.04 mm in the calibration region.

show abstract

“…Data-driven calibration for galvanometric laser scanners was proposed in [14], [15]. The authors implemented artificial neural nets (ANN), Support Vector Regression (SVR) and Gaussian processes (GP).…”

Section: Introductionmentioning

confidence: 99%

Input and Output Manifold Constrained Gaussian Process Regression for Galvanometric Setup Calibration

Boi

Sels

Moor

et al. 2022

IEEE Trans. Instrum. Meas.

View full text Add to dashboard Cite

Data-driven techniques are finding their way into the calibration procedure of galvanometric setups. However, they bypass the underlying physical or mathematical model completely. Recent work has shown that a simple assumption about an underlying truth can improve the predictions: laser beams leaving the device follow a straight line. In this paper we take that approach a step further. Both the inputs (the pairs of rotations of the two mirrors) and outputs (the straight lines) lie on a manifold. We can incorporate this prior knowledge in the model via constraints built in the formulation of a covariance function. We propose two constrained models: one in which a linear constraint on the direction vector is written as a differential equation and one in which a quadratic constraint is imposed by a reparametrization of the line coordinates. We compare them to data-driven and unconstrained model-based approaches. We show that enforcing constraints improves the quality of the predictions significantly and thus the accuracy of the calibration. We validate our findings against real world data by predicting points on validation planes, calculating line segment distances, considering the training times for the models and assessing how much a predicted line resembles an actual straight line.

show abstract

Calibration of Galvanometric Laser Scanners Using Statistical Learning Methods

Cited by 5 publications

References 5 publications

Data-Driven Learning for Calibrating Galvanometric Laser Scanners

Data-Driven Learning for Calibrating Galvanometric Laser Scanners

In situ calibration of galvanometric scanning head for laser powder bed fusion machines based on a vision system

Input and Output Manifold Constrained Gaussian Process Regression for Galvanometric Setup Calibration

Contact Info

Product

Resources

About