Home

3D Deep Learning — Solar Panel Anomaly Detection

Welcome to the project wiki. This repository implements a 3D point cloud pipeline for detecting surface anomalies in solar panels using PointNet-based deep learning.

VGG-16 CNN architecture — Conv layers extract hierarchical features, max pooling reduces spatial dimensions, fully connected layers produce the final classification. Source: LearnOpenCV

---

Wiki Pages

Page	Description
Classical CNN vs 3D CNN	Architecture differences, input representations, and when to go 3D
Why 3D Deep Learning for Solar Panels	Limitations of 2D inspection and what depth data reveals
GIS and Geospatial 3D Deep Learning	LiDAR, photogrammetry, and geospatial applications of 3D DL

Quick Links

• Repository
• PointNet Paper (Qi et al., 2017)

Interactive Tools

Tool	Description
TensorFlow Playground	In-browser neural network simulator — experiment with layers, activations, learning rate, and datasets in real time