Rachit Jaiswal
All work

CMR — LiDAR Lane Detection

In progressRobotics / Graph theory

I’m working on LiDAR lane detection using graph theory.

From returns to connected lanes

Explore the graph
Illustrative LiDAR lane graph A synthetic point cloud with two curved lane boundaries. Nearby returns form graph edges; highlighted paths trace connected boundaries. LiDAR Point cloud → graph → lane paths
Illustrative point cloud, not recorded sensor data. Select a layer and adjust the radius to explore how local connections form continuous paths.

Two illustrative lane boundaries over a synthetic point cloud.

Current work

My focus is lane detection from LiDAR using graph theory.

Explore the idea

In this illustration, each return becomes a node. Nearby nodes connect into a directed graph, and paths through that graph trace two candidate lane boundaries. A smaller connection radius can break a path; a larger radius introduces more possible connections.

The experiment demonstrates the idea of graph connectivity. It is not a reconstruction of the project’s implementation or a claim about detection accuracy.