How a satellite decides which images to send
A camera in orbit can collect more imagery than its radio link can send to Earth at once. The bottleneck is not just distance: a spacecraft has limited time in contact with ground antennas, and its communications equipment can carry only so much data during each pass. If every image must wait in a queue, useful observations may arrive late—or compete with newer ones for storage and transmission time.
An onboard screening system moves one part of that decision closer to the camera. After an image is captured, software can examine it for signs that make it less useful, such as extensive cloud cover or corrupted sensor readings. It can then flag, rank, or set aside images before they enter the transmission queue. The goal is not to replace the satellite’s camera or send an AI-generated picture; it is to decide which captured data deserves scarce communication time.
A small computer has to make a careful trade
The screening happens with limited computing power and without the full context available to people analyzing images on the ground. A system must recognize likely problems quickly, using algorithms and computing resources that fit aboard the spacecraft. A mistaken rejection could discard a valuable observation, while a missed problem could waste bandwidth on an image that offers little usable information.
That makes the filter a prioritization tool, not a perfect judge. Selected imagery and other important data still have to reach Earth, where analysts can inspect them in greater detail. By reducing unnecessary transmissions, onboard screening can leave more room for observations that are clearer or more relevant—and help turn a satellite’s brief contact with a ground station into a more useful delivery window.
