XRTool
A real-time 3D point cloud registration(matching) algorithm library for iOS, combining the 3D point cloud captured from LiDAR and movement information captured from CoreMotion. With this library you can use an iOS device with LiDAR to detect key feature points in the 3D point cloud, and the algorithm will automatically find match of them to the pre-stored CAD model. Matched feature points combination will then be used as an anchor point at which the virtual object will be placed into the AR world.
What Did I Do
- Collaborate closely with an international team. Analyze the clients requirements, design the solution
- Implement key computer vision algorithms with Apple’s frameworks such as Metal, Metal Performance Shader and CoreMotion, as well as OpenCV, Eigen3 and PointCloud
- Visualize the algorithm’s internal states and final results with ARKit and RealityKit/SceneKit/Metal
- Read, parse and render 3D model files in USDZ format with RealityKit
- Implement the UI for our prototype app with SwiftUI and UIKit combined
Tech Stack
ARKit, RealityKit/SceneKit, Metal, MPS(Metal Performance Shader), SwiftUI, PointCloud, OpenCV, Eigen3
Use the share button below if you liked it.
Any kind of appreciation, is appreciated