PopularMajor Project
Artificial Intelligence
Face Recognition Attendance System
Camera-based attendance with liveness detection and analytics
- Python
- OpenCV
- FaceNet
- Flask
- +2
₹4,900₹5,999
4.8/301 reviews
Artificial Intelligence · Capstone
LiDAR Person Detection with YOLO
Finds pedestrians in point clouds via a bird's-eye projection
5-6 weeks build
Overview
Raw LiDAR scans from the KITTI dataset are cropped, voxelised and projected into bird's-eye-view images that encode height, intensity and density as channels. A YOLO detector trained on these images locates pedestrians and cyclists, and boxes are projected back into the 3D point cloud and onto the camera image for inspection in an Open3D viewer.
What makes it stand up
- 01Point cloud to bird's-eye-view encoding with three channels
- 02KITTI labels converted into 2D training boxes
- 03YOLO trained for pedestrians, cyclists and cars
- 04Detections projected back into 3D and the camera view
- 05Average precision per class using KITTI evaluation rules
Module breakdown
- 01
Point cloud I/O
KITTI scan loading, cropping and ground filtering.
- 02
BEV encoding
Height, intensity and density maps from voxels.
- 03
Detector
YOLO training on projected images with rotated boxes.
- 04
Back-projection
Calibration maths from BEV to 3D and to camera.
- 05
Visualiser
Open3D scene with boxes and frame scrubbing.
After this, you will be able to
- Explain why BEV projection makes LiDAR usable by 2D detectors
- Walk through the KITTI calibration matrices
- Demo detections rendered in a 3D point cloud
Technology stack
- Python
- PyTorch
- YOLOv8
- Open3D
- NumPy
- Level
- Capstone
- Domain
- Artificial Intelligence
- Build time
- 5-6 weeks
- Difficulty
You receive
- Complete, commented source code
- 60-90 page project report (IEEE format)
- Editable presentation deck
- Architecture, ER and UML diagrams
- Local setup and deployment guide
- Sample dataset or seed data
- Viva question bank with answers
- One 45-minute walkthrough call
- 30 days of doubt-clearing support