- Recycle Image Set
- How to Train Tensorflow Object Detection Model on Windows 10
- How to Train Tensorflow Lite Object Detection Model in Cloud TPUs
- How to Build and Install Tensorflow From Source on Windows 10
- Object Detection with Google Coral USB Accelerator
- USB-to-Serial Communication with Raspberry Pi and Arduino
- 3D Printed Robotic Arm
Arduino - files used for Arduino to control robotic arm:
- basicMovement - .ino file to perform basic movement with robotic arm
- roboticArm - .ino file to control robotic arm to pick up and drop off a piece of recycling detected from the Raspberry Pi
Documentation - files documenting my build process:
- Bill_of_Materials.pdf - .pdf file including all my parts used in my project
- Notebook.pdf - .pdf file including my notebook that I used when building my robot
Images - images of my project
RPI - files used for Raspberry Pi to perform object detection:
- recycle_ssd_mobilenet_v2_quantized_300x300_coco_2019_01_03 - .tflite and .txt files of the trained recycle model and lables to perform object detection
- recycle_detection.py - python code used to perform object detection and communicate with Arduino
Tensorflow - files used for training model:
- CSV - .csv files of both the train and test image set
- Images - .jpg images of recycling materials and .xml files of the annotated images
- SaveModel - frozen graph of trained model
- TFLite - .tflite file of trained model and .tflite file of compiled model for Google Coral USB Accelerator
- TFRecord - .record files of both the train and test images set used for training
- ssd_mobilenet_v2_quantized_300x300_coco_2019_01_03 - .config, .ckpt, and .pb files of the quantized model used for transfer learning
- labelmap.pbtxt - label map used for training