Clone this repo and cd into to it:

git clone https://github.com/frc2423/tensorflow-object-detection-example.git cd tensorflow-object-detection-example 

From the tensorflow-object-detection-example directory run these commands:

git clone https://github.com/tensorflow/models.git docker build -f ./Dockerfile -t kwarqs-object-detection . 

docker run --privileged --device=/dev/video0:/dev/video0 -it kwarqs-object-detection 

To access the running container first get its name by running docker ps in the terminal:

And then run docker exec -it -p 6006:6006 [container name] /bin/bash:

To develop in vscode first install the docker extension:

To update the docker image, right click the Dockerfile in vscode and click Build Image...:

Tag the image as kwarqs-object-detection:latest and press enter:

To run the image right the kwarqs-object-detection:latest image in the docker vscode extension and click Run Interactive:

This should launch a shell in vscode into the container:

To launch another shell, stop and start the container, and perform other actions right click the container in vscode:

From the kwarqs directory inside the container generate the TFRecords:

python generate_tfrecord.py --csv_input=images/train_labels.csv --image_dir=images/train --output_path=training/content/train.record python generate_tfrecord.py --csv_input=images/test_labels.csv --image_dir=images/test --output_path=training/content/test.record 

To train the model, execute the following command in the command line:

python ../models/research/object_detection/model_main_tf2.py --pipeline_config_path=training/ssd_efficientdet_d0_512x512_coco17_tpu-8.config --model_dir=training --alsologtostderr 

Run tensorboard to visualize training of the model:

tensorboard --logdir=training/train --host 0.0.0.0 --port 6006 

When you are done training, run this to generate the inference graph that can be used to run the model:

python ../models/research/object_detection/exporter_main_v2.py --trained_checkpoint_dir training --output_directory inference_graph --pipeline_config_path training/ssd_efficientdet_d0_512x512_coco17_tpu-8.config 

To detect objects with your webcam using the model you trained run this script:

python detect_from_webcam.py --labelmap training/content/labelmap.pbtxt --model inference_graph/saved_model