# The First Moon Landing Wasn’t Apollo — And We Just Found It Source: https://www.youtube.com/watch?v=rrRauDKoTJM Recap page: https://rapidrecap.app/video/rrRauDKoTJM Generated: 2026-02-28T16:33:28.849+00:00 --- ## Quick Overview A recent study utilizing a novel machine learning algorithm (YOLO-ETA) suggests the possible re-identification of the Soviet Luna 9 landing site, which occurred in 1966, by analyzing high-resolution Lunar Reconnaissance Orbiter Camera (LROC) imagery and finding high-confidence matches for discarded spacecraft hardware, contrasting with the fact that the US Apollo missions' landing sites are already confirmed. **Key Points:** - A research group used a novel machine learning algorithm, YOLO-ETA (You-Only-Look-Once – Extraterrestrial Artefact), adapted from TinyYOLOv2, to search LROC imagery for anthropogenic objects on the Moon. - The algorithm was trained on Apollo landing site data and achieved a balanced precision-recall (F1 = 0.60) and an 80% mean confidence score for lander detections in previously unseen images. - Applying the model to a 5x5 km region surrounding the historically uncertain Luna 9 landing area yielded several high-confidence detections of artificial objects near 7.03° N, -64.33° E. - The analysis indicates that the candidate site's horizon geometry is potentially consistent with Luna 9's topographical panoramas, suggesting a possible location for the first Soviet Moon landing in 1966. - The researchers also successfully detected hardware remnants from the Apollo missions (like the Lunar Module, flag, LRRR, PSEP) and the Soviet Luna 16 lander, demonstrating the system's effectiveness. - The study suggests this AI-driven methodology can support future orbital surveys of lunar artifacts and surface details, potentially resolving ambiguities in historical landing locations. ![Screenshot at 0:05: The title card and an image of the Soviet Luna 9 replica, highlighting the subject of the investigation—the first successful soft landing on the Moon.](https://ss.rapidrecap.app/screens/rrRauDKoTJM/00-00-05.jpg) **Context:** The video discusses a recent scientific paper published in npj Space Exploration (2026) detailing the potential identification of the landing site for the Soviet Luna 9 probe, the first spacecraft to achieve a soft landing on the Moon in 1966. This contrasts with the US Apollo landing sites, which are already well-documented and imaged by NASA's Lunar Reconnaissance Orbiter (LRO). The key development is the application of a custom machine learning system called YOLO-ETA to search historical LROC images for small, artificial debris left behind by these early missions. ## Detailed Analysis The video reports on a breakthrough in lunar archaeology: the possible identification of the Soviet Luna 9 landing site using artificial intelligence. The first Moon landing was not Apollo, but the uncrewed Soviet Luna 9 mission in February 1966, three years before Apollo 11. Since then, confirming the precise location has been difficult. A group of researchers developed a lightweight computer vision system called YOLO-ETA (You-Only-Look-Once – Extraterrestrial Artefact), adapted from TinyYOLOv2, to detect anthropogenic objects in high-resolution LROC imagery. Trained on Apollo landing site data, the system achieved a decent F1 score (0.60) and could correctly localize known hardware like the Lunar Module (LM), flag, and various instruments from Apollo missions (e.g., Apollo 11, 12, 14, 15, 16). When applied to the historically uncertain region around the expected Luna 9 landing area, the AI identified several high-confidence anomalies that align with the expected location (7.03° N, -64.33° E). The topographical analysis of the candidate site's horizon geometry is consistent with the panorama photos sent back by Luna 9, suggesting the AI successfully pinpointed the crash site debris, including discarded components. The researchers also successfully located the Luna 16 lander and Surveyor 7 hardware. The speaker emphasizes that this advancement in image analysis technology is crucial for future lunar exploration and for locating all historical space hardware, noting that the US astronauts left behind artifacts like the plaque on the Eagle lander, whereas the Soviets often left debris, and that China is actively pursuing a lunar base on the North Pole. ### Luna 9 Identification - A group of researchers used a novel machine learning algorithm, YOLO-ETA, trained on Apollo data, to scan LROC imagery for artificial objects near the historically uncertain Luna 9 landing site (7.03° N, -64.33° E). - The AI found several high-confidence detections whose horizon geometry matches Luna 9's original panorama photos, strongly suggesting the landing location has been found. - The successful detection rate (80% mean confidence score for landers) suggests this method is viable for finding other historical artifacts. ### AI Validation and Other Detections - The YOLO-ETA system was validated by successfully detecting hardware from multiple Apollo missions (LM, LRRR, PSEP) and the Soviet Luna 16 lander. - The AI demonstrated an ability to identify discarded components and even features in areas with unfavorable lighting conditions. - The success rate for Luna 9 (43.8% confidence) was lower than for Apollo hardware (up to 78.4% for some components) but still significant. ### Implications for Future Exploration - The ability to automatically map and identify hardware relics is critical for future lunar missions, especially for establishing permanent bases near the North Pole, as China is planning. - The speaker contrasts the US focus on displaying plaques with the Soviet practice of leaving debris, implying a need to catalog all human impact on the Moon. - The video concludes with a call to action to use Brilliant.org to learn more about related topics like AI, probability, and digital circuits. ![Screenshot at 0:00: Sabine Hossenfelder introduces the topic, displaying the title of the research paper regarding the possible identification of the Luna 9 landing site using machine learning.](https://ss.rapidrecap.app/screens/rrRauDKoTJM/00-00-00.jpg) ![Screenshot at 0:14: A replica of the Soviet Luna 9 probe, which achieved the first soft landing on the Moon in 1966, is shown.](https://ss.rapidrecap.app/screens/rrRauDKoTJM/00-00-14.jpg) ![Screenshot at 0:27: The first ever photograph taken from the surface of the Moon by Luna 9, showing the heavily cratered terrain.](https://ss.rapidrecap.app/screens/rrRauDKoTJM/00-00-27.jpg) ![Screenshot at 1:19: A high-resolution LROC image of the Apollo 11 landing site, clearly labeling the Lunar Module \(LM\), Camera, LRRR, PSEP, and Discarded Cover.](https://ss.rapidrecap.app/screens/rrRauDKoTJM/00-01-19.jpg) ![Screenshot at 4:44: Close-up of the plaque left on the Apollo 11 Lunar Module descent stage, which states the mission came in peace for all mankind.](https://ss.rapidrecap.app/screens/rrRauDKoTJM/00-04-44.jpg)