Back

Comments 11

  • Recycle Sorting Robot With Google Coral about 1 year ago

    Thank you for your question. What I did, was connect the RPI to an Arduino via a USB cable. You will have to follow the instructions here: https://www.sunfounder.com/blog/rpi-ard/. It will show you how you can have the RPI communicate with an Arduino via serial communication. So what I did was depending on where the detected object was located, the RPI would send a message to the Arduino to move the servo motors at certain angles in order to pick up the object. I hope this answered you question. Let me know if you need further help.

  • Recycle Sorting Robot With Google Coral about 1 year ago

    Hmm... It seems like you don't have the Edge TPU Python API installed. Try running this sudo apt-get install python3-edgetpu
    . Let me know if that works.

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Thank you for letting me know.

    Have you completed this tutorial: https://coral.ai/docs/accelerator/get-started/#on-linux

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Sent you an email!

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Glad to help!

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Yes, run the all the computer vision code on the rpi with the edge tpu connected.

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Sorry for miss clarification. If you want to train you own model with the dataset-resized in the google drive, follow steps 1 to 4. I think we are mixing up which pre-trained models we are referring to. To train your own model, you can use a edge tpu compatible pre-trained model for transfer learning as seen in step 3. If you want to use a model I already trained on the dataset, you can see the "Recycle Detection Model" section in the tutorial. Let me know if I answered your question.

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Yes, all you have to do is clone my github repo and run the pre-trained model with this command: python test_detection.py --model recycle_ssd_mobilenet_v2_quantized_300x300_coco_2019_01_03/detect_edgetpu.tflite --labels recycle_ssd_mobilenet_v2_quantized_300x300_coco_2019_01_03/labels.txt

    I am not sure what you mean by the other object detection models in you second question.

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Thank you! To use the pre-trained model, you will have to setup a raspberry pi and connect the camera module & coral edge tpu. Please follow step 5 in the tutorial. To use the arduino and raspberry pi together, you need to connect a usb cable from the arduino and rpi usb port. This allows serial communication between each other. I have updated step 7 in the tutorial that refers to a link explaining how this works.

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Thanks! I don't have a quantifiable measurement for how tightly the arm grips an object. But it does depend on the object size. It can only hold objects that can fit in the gripper.

  • Recycle Sorting Robot With Google Coral over 1 year ago

    Sent you an email!

Add projectSign up / Login