Related Experiment Video
Updated: Oct 22, 2026

Bringing the Visible Universe into Focus with Robo-AO
Published on: February 12, 2013
Defocus estimation for star sensors based on closed energy contours
Hongzhe Chen1,2, Wenbo Zhang1,2,3, Weiguo Lu1,2
1Chinese Academy of Sciences, Xi'an Institute of Optics and Precision Mechanics, Xi'an 710119, China.
Abstract:
This paper presents a novel, to our knowledge, defocus estimation method for star sensors, leveraging the equivalent radius R80 of the 80%-energy closed contour in conjunction with pixel-phase compensation. In contrast to existing approaches, the proposed method operates on a single isolated-star-spot image and circumvents the need for complex parametric reconstruction of the point-spread function. The processing pipeline comprises background subtraction, noise suppression, centroid estimation, and pixel-phase compensation, followed by subpixel interpolation and threshold-based searching to extract the largest connected region that encloses approximately 80% of the total spot energy. The radius of an equal-area circle is then adopted as R80, and a quadratic calibration model is established to relate R80 to the axial defocus. Experiments conducted over a defocus range from -0.46875 to 0.46875 µm confirm that R80 varies quadratically with defocus. Notably, pixel-phase compensation reduces the mean standard deviation of R80 from 0.0426 to 0.0214 pixels, yielding a 49.8% improvement in repeatability. On independently acquired test images, the mean absolute bias of the seven position-wise mean estimates is 0.0092 µm. Owing to its single-image requirement and avoidance of iterative PSF fitting, the method is particularly well suited for practical focal-plane alignment and on-orbit image-quality assessment in star sensor systems.
Related Concept Videos
Focusing of Light in the Eye
Difference from Background: Limit of Detection
The LOD indicates the presence or absence...
Plotting and Calibrating the Root Locus
The maximum gain occurs at the breakaway points between open-loop poles on the real axis, while the minimum gain is observed...

