What corner-seeing iPhone LiDAR technology means
Corner-seeing iPhone LiDAR technology is a form of non-line-of-sight imaging that uses the iPhone LiDAR sensor and motion to infer the presence, position, and rough shape of objects hidden beyond direct view, turning scattered, noisy depth data into practical object detection iPhone systems that can warn users about activity around corners. This new corner detection technology builds on LiDAR, a laser-based depth sensor already built into recent iPhone Pro and iPad Pro models, but pushes it far beyond simple room scanning. MIT LiDAR research from the Media Lab shows that consumer-grade sensors, including those that cost under USD 50 (approx. RM230), can detect and track objects that sit completely outside the camera’s field of view. That step moves non-line-of-sight imaging from lab-only demonstrations toward tools that could one day run on everyday phones and tablets.
How MIT made an iPhone-class LiDAR see around corners
To understand how this works, picture light bouncing off a nearby wall, reflecting onto a hidden object, then scattering back to the wall and into the iPhone LiDAR sensor. Each individual LiDAR scan is noisy and incomplete, but the system collects many readings as the phone moves, treating each camera position like a tiny piece of a larger virtual aperture. The researchers call this an aperture sampling model that stitches those imperfect depth samples into a single, more coherent picture over time. Instead of a sharp image, the algorithm builds a probabilistic model that reveals whether something is there, how it moves, and its approximate shape. According to MIT Media Lab, the same low-power, consumer-grade LiDAR already inside smartphones is enough to support this non-line-of-sight reconstruction, as long as software can access raw depth data and motion information from the device.
From hidden blobs to meaningful object detection
The system’s results look more like ghostly blobs than photo-realistic reconstructions, but they contain vital information. As the phone shifts position, the aperture sampling model refines its estimates of where a hidden object sits and how it moves relative to the camera. Over several frames, the non-line-of-sight algorithm can track a single object’s trajectory, infer its rough outline, and even distinguish between multiple moving objects tucked around the corner. That makes the iPhone LiDAR sensor behave a bit like echolocation with light instead of sound. The focus is not on pretty images but on reliable corner detection technology that answers questions such as: Is there something there? Is it moving toward or away from me? Is it large or small? For many safety and navigation tasks, those answers are more than enough.
Why this matters for iPhones, robots, and AR
The MIT team demonstrated four headline abilities: tracking a single hidden object, reconstructing its shape, tracking multiple hidden objects at once, and using obscured landmarks for camera self-localization. That last feature is especially promising for robots, drones, and future self-driving systems: a machine that can orient itself using things it cannot directly see can avoid collisions and plan smarter paths. For everyday users, the same principles could power object detection iPhone apps that warn about people or vehicles approaching an intersection, aid low-vision navigation, or add more convincing augmented reality interactions that respect hidden obstacles. While you cannot yet run this on an iPhone, the researchers have released their code publicly, and similar LiDAR hardware can be built for under USD 50 (approx. RM230), which opens the door for developers and device makers to experiment now.