# Computationally intensive technique to search for TNO's within the noise of historical photographic data?

I remember reading recently about a technique to search through lots of old survey images to try to identify trans-Neptunian objects that were so dim that they were "lost in the noise" in single exposures and so have here-to-fore escaped detection.

This was a few months ago I think, but now I can not remember where or exactly when.

From what I remember it was a computationally intensive technique where images were effectively overlayed with offsets determined by hypothetical test orbits. In other words suppose a field is photographed one a year for 10 years. 1 or 2 sigma noise-like dots of diffraction-limited size will be ignored in any one image, but if many of them show dots that overlap when images are shifted by 0.126 degrees per year, then this may be due to a real TNO at 200 AU. ($$\text{1 year} \times (\text{180} / \pi) \times \sqrt{GM/a^3}$$)

Question: What was this computationally intensive technique to search for TNO's within the noise of historical photographic data that I'm remembering? Where was it recently published or at least written about?

Dark Energy Survey data have yielded several TNO discoveries. Bernardinelli et al. 2020 describe their method in detail. First they compute a trial orbit from images where a transient object was detected. Then

Final validation of the reality of linked orbits uses a new "sub-threshold confirmation" test, wherein we demand the object be detectable in a stack of the exposures in which the orbit indicates an object should be present, but was not individually detected.

Their figure 10 shows a candidate object passing that test; figure 11 shows one failing.

Rice and Laughlin 2021 use a baseline subtraction and shift-stacking technique to search TESS data for TNOs without prior detections. After verifying that they recover a few known TNOs this way, they list a few unknown candidates for further observation.