In this project, I perform object detection on the Raspberry Pi using the Google Coral USB Accelerator. A mobilenet_ssd _v2_coco_quant_postproces_edgetpu.tflite TensorFlow lite model from Google is used in my project. This project is modified as a security camera, filming a 15-second video and sending a text message via Pushetta when a pre-set object from the coco_labels.txt is detected and logging the objects detected every second to a text file. Both the video and text file will be uploaded to Google Drive.
https://bluestampengineering.com/student-projects/taiga-a/
~/python-tflite-source/edgetpu/demo
caller.py
detectVideoMod.py
record.py
log.py
~/edgetpu_models/
mobilenet_ssd_v2_coco_quant_postprocess_edgetpu.tflite
coco_labels.txt