What ‘seeing around corners’ with iPhone LiDAR really means
Corner detection technology using the iPhone LiDAR sensor is a form of non-line-of-sight imaging that infers the presence, movement, and rough shape of hidden objects by analyzing scattered light as the phone moves, effectively letting the device detect things outside the camera’s direct view without adding any new hardware. In practice, this does not create a sharp photo of what is hiding behind a wall. Instead, the system builds a coarse but useful model of nearby hidden objects over time. MIT Media Lab researchers proved that consumer-grade LiDAR, similar to the sensor in iPhones and iPad Pro, can detect and track objects that are entirely outside the camera’s field of view. Their work shows that non-line-of-sight imaging no longer belongs only in specialized labs; it can live in everyday augmented reality hardware you already carry.
How corner detection technology works on existing hardware
The iPhone LiDAR sensor sends out pulses of light and measures how long they take to bounce back, building a depth map of the scene. For non-line-of-sight imaging, MIT’s system watches how this depth data changes as you move your phone, using motion as the “secret sauce” to infer what is happening around the corner. The researchers call their method an aperture sampling model. It stitches together many noisy, partial readings from slightly different viewpoints into a single coherent estimate of hidden objects. Over a short period, the system can tell that something is present, guess its approximate shape, and track how it moves, much like echolocation with light instead of sound. According to MIT Media Lab, the same low-power LiDAR already embedded in smartphones is enough to support this style of computational imaging without any new sensors.
From lab demo to smartphone computational imaging
Non-line-of-sight imaging used to depend on powerful, expensive lab-grade lasers, which kept it far away from consumer devices. The new research changes that by showing similar effects with low-cost LiDAR hardware. The sensor setup used by the MIT team can be assembled for under USD 50 (approx. RM230), yet it still detects and tracks hidden objects. Even more important, the code for the system is publicly available, so developers and researchers can experiment and build on it. The main limitation today is access to raw LiDAR data on commercial phones, which platform owners do not always provide. Still, the work proves that existing augmented reality hardware already has the physical capabilities needed. The next leap in smartphone computational imaging may arrive through software updates and APIs, not new camera modules.
New photography, AR, and gaming tricks for iPhone
If corner detection technology reaches future iOS releases or third-party apps, it could reshape how people use the iPhone LiDAR sensor every day. For photography, hidden subject awareness could prevent focus failures when someone is about to walk into frame, or help auto-adjust exposure as a person emerges from behind a wall. In augmented reality, games could react to characters or objects before you see them, adding tension and surprise based on what lies around the corner. AR navigation might warn that a corridor is blocked or that someone is approaching from a blind spot, smoothing movement in crowded areas. These features depend on computational imaging: software that interprets raw depth data in smarter ways, so the phone understands more about the scene than the camera alone can see.
Safety, robotics, and what comes next
Beyond consumer apps, non-line-of-sight sensing with phone-class LiDAR opens new paths for safety and robotics. The MIT team showed four core capabilities: following a single hidden object, reconstructing its shape, tracking multiple hidden objects, and using unseen landmarks for camera self-localization. That final skill is especially helpful for robots, self-driving systems, and delivery drones that need to orient themselves in complex environments with limited visibility. A future phone or headset could quietly watch for motion around blind corners and alert you before you bump into someone, or give early warning about obstacles when your view is blocked. For developers, the combination of public code and widely available hardware lowers the barrier to experimentation. The next generation of augmented reality hardware may focus less on adding sensors and more on teaching existing ones to see the unseen.