Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Appearance settings

using a 3D Model for training your Object Detection (Yolo, CustomVision)

Notifications You must be signed in to change notification settings

uneidel/3dModelObjectDetection

Open more actions menu

Repository files navigation

Prerquisites

Blender - tested with 2.79 - www.blender.org
VOTT - https://github.com/Microsoft/VoTT
Download actual build from https://github.com/Microsoft/VoTT for a Visual Tagging Tool
pip install azure-cognitiveservices-vision-customvision

Blender Activties

Add Stl to empty Blender or using py Function CreateFromScratch(pathToStl,"c:\temp\render.blend") to create new Scene

(
## Set Active Camera
Move Active Camera to View
Hotkey: Ctrl-Alt-Numpad0
# Check with
Hotkey: Numpad0
)

#Run Script
path "c:\Program Files\Blender Foundation\Blender\2.79"
#Run via blender to ease the use of bpy module
blender --background -P render.py

Process

alt text

After the Process you receive a json file with metadata including Boundingboxes and the rendered pictures

Object Detection Yolo

alt text

USE Azure DataScienceVM to train the model:

Setup GPU Machine for Darknet Training

Install Cuda on 18.04

wget https://developer.nvidia.com/compute/cuda/9.2/Prod2/local_installers/cuda_9.2.148_396.37_linux
wget https://developer.nvidia.com/compute/cuda/9.2/Prod2/patches/1/cuda_9.2.148.1_linux
sudo apt install -y make gcc freeglut3 freeglut3-dev libxi-dev libxmu-dev
Sudo sh cuda_9.2.148_396.37_linux

Follow the instructions
- accept Eula
- agree to non unspported configuration
- agree to NVIDIA Accelerated Graphics Driver
- agree to OPENGL Driver
- disagree to nvidia-xconfig
- agree to CUDA 9.2 Toolkit (use default for the subsequent)
- agree to install the Samples

Install patch*
Sudo sh cuda_9.2.148.1_linux

''Testing Installation**
Change to deviceQuery Folder:
cd samples/1_Utilities/deviceQuery/
make
Query for existing CUDA enabled Cards:
./devicequery

Install Darknet and Compile for GPU

git clone https://github.com/pjreddie/darknet
edit Makefile
and change GPU=0 to GPU=1 (DEGUG=0 to DEBUG=1)
make

test installation

./darknet detect cfg/yolov3.cfg yolov3.weights data/dog.jpg

#Train the model [!TODO]

Build and Test

Custom Vision

Upload to CustomVision

./python VOTTtoCV.py

Export to customvision

alt text

Score with Customvision

alt text

About

using a 3D Model for training your Object Detection (Yolo, CustomVision)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

Morty Proxy This is a proxified and sanitized view of the page, visit original site.