In Situ Enhancement of Heliostat Calibration Using Differentiable Ray Tracing and Artificial Intelligence




Heliostat Calibration, Differentiable Raytracing, Machine Learning


The camera target method is the most commonly used calibration method for heliostats at solar tower power plants to minimize their sun tracking errors. In this method, individual heliostats are moved to a white surface and their deviation from the targeted position is measured. A regression is used to calculate errors in a geometry model from the tabular data obtained in this way. For modern aim point strategies, or simply heliostats in the rearmost end of the field, extremely high accuracies are needed, which can only be achieved by many degrees of freedom in the geometry model. The problem here is that the camera target method produces only a very small data set per heliostat, which limits the number of free variables and thus the accuracy. In this work, we extend existing ray tracing methods for solar towers with a differentiable description, allowing for the first time a data-driven optimization of object parameters within the ray tracing environment. Therefore, the heliostat calibration can take place directly within the ray tracing environment. Thus, the image data acquired during the measurement can be processed directly and more information about the orientation of the heliostat can be obtained. Within a simple example we show the advantages of the method, which converges faster and corrects errors that could not be considered before. Without any disadvantages or additional costs, the state-of-the-art calibration method can be improved.


Download data is not yet available.


C. Sattler et al. Review of heliostat calibration and tracking control methods, Solar Energy, Vol. 207, Elsevier, p. 110-132, 2020 DOI:

B. Nicolet et al. “Large steps in inverse rendering of geometry” ACM Transactions on Graphics (TOG), Vol. 40, No. 6, p. 1-13, 2021 DOI: 10.1145/3478513.3480501

G. Loubet et al. “Reparameterizing discontinuous integrands for differentiable rendering” ACM Transactions on Graphics (TOG), Vol. 38, No. 6, p. 1-14, 2019 DOI: 10.1145/3355089.3356510

T. Li et al. “Differentiable monte carlo ray tracing through edge sampling”, ACM Transactions on Graphics (TOG), Vol. 37, No. 6, p. 1-11, 2018 DOI: 10.1145/3272127.3275109

S. Pattanaik et al. “Computation of global illumination by Monte Carlo simulation of the particle model of light”, Third Eurographics Workshop on Rendering, p.71-83 1992 DOI: 10.1002/vis.4340040303

H. Kato et al. “Differentiable rendering: A survey”, arXiv preprint arXiv:2006.12057, 2020 DOI:

Q. Sun et al. “End-to-end complex lens design with differentiable ray tracing”, ACM Transactions on Graphics, Vol. 40, No. 4, p. 1-13, 2021 DOI: 10.1145/3450626.3459674

L. Jain et al. “Generating Semantic Adversarial Examples with Differentiable Rendering”, CoRR, Vol. abs/1910.00727, 2019 DOI:

J. Tu et al. “Exploring Adversarial Robustness of Multi-Sensor Perception Systems in Self Driving”, arXiv 2021, doi:

Y. He, “Numerical simulation of solar radiation transmission process for the solar tower power plant: from the heliostat field to the pressurized volumetric receiver”, Applied thermal engineering, Vol. 61, No. 2, p. 583-595, 2013 DOI:

K. Stone Automatic heliostat track alignment method, Google Patents, 1986

R. Rojas, “The backpropagation algorithm” Neural Networks, Springer, p. 149-182, 1996 DOI: 10.1007/978-3-642-61068-4_7

H. Yu “Levenberg-marquardt training”, Industrial electronics handbook, Vol. 5, No. 12, p.1, 2011 ISBN: 9781315218427

A. Suratgar “Modified Levenberg-Marquardt method for neural networks training”, World Acad Sci Eng Technol, Vol. 6, No. 1, p.46-48, 2005

A. Reynaldi, “Backpropagation and Levenberg-Marquardt algorithm for training finite element neural network”, Sixth UKSim/AMSS European Symposium on Computer Modeling and Simulation, p. 89-94, 2012 DOI: 10.1109/EMS.2012.56

F. Schroff, “Facenet: A unified embedding for face recognition and clustering”, Proceedings of the IEEE conference on computer vision and pattern recognition, p 815-823, 2015 DOI: arXiv:1503.03832

W. Chen, “Beyond triplet loss: a deep quadruplet network for person re-identification”, Proceedings of the IEEE conference on computer vision and pattern recognition, p. 403-412, 2017 DOI: arXiv:1704.01719

J. Ribera, “Weighted hausdorff distance: A loss function for object localization”, arXiv preprint arXiv:1806.07564, Vol. 2, p.1, 2018 DOI: arXiv:1806.07564s




How to Cite

Pargmann, M., Ebert, J., Kesselheim, S., Maldonado Quinto, D., & Pitz-Paal, R. (2024). In Situ Enhancement of Heliostat Calibration Using Differentiable Ray Tracing and Artificial Intelligence. SolarPACES Conference Proceedings, 1.

Conference Proceedings Volume


Analysis and Simulation of CSP and Hybridized Systems