ratatouille-robotics / ingredient_validation

This repository contains the ROS Node to validate the ingredient and publish the information into a topic.

Ingredient Validation Pipeline

This repository contains the ROS Service that validates the ingredient using an RGB image. If the ingredient is visually ambiguous (such as salt, sugar, oregano or black pepper), it is validated using its spectral absorbance reading. Finally, the identified ingredient name is sent as a response to the caller.

To run the service:

rosrun ingredient_validation validate_ingredient_service.py

To call the service via a client:

1. RGB Mode:

rospy.wait_for_service('ingredient_validation')
service_call = rospy.ServiceProxy('ingredient_validation', ValidateIngredient)
response = service_call(mode='rgb')

2. Spectral Mode:

rospy.wait_for_service('ingredient_validation')
service_call = rospy.ServiceProxy('ingredient_validation', ValidateIngredient)
response = service_call(mode='spectral', ingredient_name="no_ingredient")

To run data collection:

1. Create a folder named data in ingredient_validation package.
2. Modify line 40 in the script with the name of the ingredient whose image data is being collected.
3. Run the script with the command: rosrun ingredient_validation data_collection.py
4. Type s to draw the bounding box around the object that has to be tracked across frames and press enter.