Processing

Mosaicking

Stitches many scenes into a seamless, cloud-removed composite over a large AOI.

INPUT SCENESscene A · 2025-03-12scene B · 2025-03-20scene C · 2025-03-28MOSAICcloud mask+ seam pathCOMPOSITE OUTPUTABCseam-line path follows low-contrast features
Fig. 1 Mosaicking composites overlapping scenes into a single seamless output over the AOI. Cloud-contaminated regions are masked out and replaced from other dates in the stack; seam-lines are chosen to cut along low-contrast features so the join is invisible.

A mosaic is a single seamless image built from tens to thousands of individual scenes captured on different dates, at different sun angles, under different atmospheric conditions. Done well, the result looks like a single image taken at one moment over a continent.

Why most shops can't do it well

You are picking the best pixel for each location. Masking clouds and cloud shadows — including thin cirrus, which rule-based detectors routinely miss. Choosing seam lines that do not cut through features a human would notice. Adjusting for sun angle differences across the AOI. Geopera's mosaics use custom path-finding for seam selection and machine learning cloud masking trained on the volume of data we process. The result is an image that looks like it was taken in a single moment.

Try it yourself test

Try doing this yourself and you will hire three people for six months. We do it by default for any imagery order over a multi-scene AOI.