IEEE SENSORS JOURNAL, VOL. 18, NO. 4, FEBRUARY 15, 2018 1449 Optimization of Temperature, Targets, and Illumination for High Precision Photogrammetric Measurements Louise Dauvin , Holger Drass, Leonardo Vanzi, Rolando Dünner, Miguel Torres, Clémentine Béchet, David Boettger, Felipe Rojas, and Tzu-Chiang Shen Abstract— The cameras of a close-range photogrammetry system must be calibrated to find their positions and optical properties. This is a crucial step in ensuring the performance of the complete system, especially if micrometric precision over a large field of view and long-term reproducibility is required. The target positions within the images for camera calibration and measurements are perturbed if stability between images is not achieved. This paper aims to evaluate the optimal conditions to guarantee the stability of the target images during measurements and calibration. The effect of temperature on the target positions is estimated to be about 0.1 pixels during a camera warm- up period of 20 min or more. This effect is reduced to about 0.02 pixels after a warm-up period of less than 10 min by developing a controller for the camera’s pixel clock. Additionally, an approximately 22% variation of illumination intensity is found to cause a small, but clearly measurable effect between images. Other aspects studied here are the illumination angle and the target characteristics. A 180 ◦ movement of the illumination source with respect to the camera produces a 0.3 pixel change in the target locations when masked retroreflective targets are involved. In contrast, this effect is diminished by a factor of about ten with the use of larger, opaque targets. Index Terms— Close-range photogrammetry, high-precision camera calibration, temperature, illumination, calibration targets, room conditions. Manuscript received October 2, 2017; revised November 17, 2017; accepted November 21, 2017. Date of publication November 27, 2017; date of current version January 18, 2018. This work was supported in part by PIA Anillo under Grant ACT-1417 and in part by FONDECYT Regular under Grant 1171364. The work of L. Dauvin was supported by Premio Padre Hurtado PUC. The work of H. Drass was supported by the FONDECYT Project under Grant 3150314. The work of D. Boettger was supported by the FONDECYT Project under Grant 3150504. This paper was presented as part of the M.S. thesis at the Pontificia Universidad Católica de Chile, June 2017. The associate editor coordinating the review of this paper and approving it for publication was Prof. Aime Lay-Ekuakille. (Corresponding author: Holger Drass.) L. Dauvin, L. Vanzi, and M. Torres are with the Department of Electrical Engineering and the Center of Astro-Engineering UC, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile (e-mail: lcdauvin@uc.cl). H. Drass is with the Department of Electrical Engineering and the Center of Astro-Engineering UC, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile, and also with the Millennium Institute of Astrophysics, Santiago, Chile (e-mail: hdrass@aiuc.puc.cl). R. Dünner and D. Boettger are with the Faculty of Physics, Institute of Astrophysics and the Center of Astro-Engineering UC, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile. C. Béchet is with the Institute of Mathematical and Computational Engineer- ing and the Center of Astro-Engineering UC, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile. F. Rojas is with the Department of Computer Science, and the Center of Astro-Engineering UC, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile. T.-C. Shen is with BlueShadows Ltda., Santiago 7500521, Chile, and also with the Center of Astro-Engineering UC, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile. Digital Object Identifier 10.1109/JSEN.2017.2777940 I. I NTRODUCTION P HOTOGRAMMETRY is the technique of making metrol- ogy measurements based on images. This technique is used in several applications, such as industrial or manufactur- ing processes, medical instruments, scientific measurements, and geology or topographic mapping. Some applications require very precise measurements, specially in close-range photogrammetry systems. An example is the positioning of small robotic arms in a spectrograph for astronomical pur- poses, which is the case study that motivated this paper. For more details see [1] and [2]. Besides, being very precise, being cost efficient becomes increasingly important to scale-up this kind of systems. Inasmuch as the position measurements obtained from pho- togrammetry are calculated using only the information given by one camera or a set of cameras, the measurement conditions and the camera properties are crucial to the performance of the complete system [3]. These properties are estimated as a part of the “Camera Calibration” process, which con- strains the ultimate performance and accuracy of the complete system [3]–[10], thus, playing a vital role in high precision applications. The camera calibration is the procedure of relating the metric information of the World Coordinate System (WCS) to its projection in the Image Coordinate System (ICS) [7], as seen in Fig.1. Both extrinsic and intrinsic parameters are calculated and optimized in order to reduce the projection errors. While the extrinsic parameters are defined as the position and orientation of the camera in the WCS, the intrinsic parameters describe the camera’s optical characteristics, such as the effective focal length, the scale factor, and the center point where image plane and optical axes intersect. Lens distortions in both radial and tangential directions are also measured as part of the camera calibration, as well as the affinity and skewness of the CCD pixels [4], [6], [11]–[13]. In this research, the images for calibration and measure- ments are taken of easily detectable targets with a well known location. The target positions are found with photometry- based centroiding in the ICS. The measured centroids of the targets are directly affected by the image quality. This requires working within the camera sensor’s linear range, using a stable camera focus, and having a homogeneously illuminated Field Of View (FOV). Performing the image acquisition in optimal conditions is necessary to have precise target positioning in the images for photogrammetric measurements and camera 1558-1748 © 2017 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.