scvf 1.1.2

a framework that wraps opencv for our needs in FRC

Spikes ComputerVision Framework

Spikes ComputerVision Framework, or scvf for short, is a framework that is made to make programming CV for the FRC easier

This framework wraps opencv and pipelines generated by grip in a way that is efficient and elegant.

Installation Instructions:

$ python3 -m pip install scvf

API

communication through network tables

use the pipeline_name key to send the pipeline_name
use the camera_id key to send the camera id
use the exposure key to send the exposure for the camera

pipeline compatibility

Though we recommend grip as the main tool to generate cv2 pipelines using grip is not strictly required
You can provide any object to server as a pipeline as long as it contains the next two methods:
process() - a method that processes a given image
get_output() - a method that returns the output of the processing

IO functions

scvf receives two functions that are responsible for communications with external data sources.

1. settings_supplier(callback) this function receives settings and supplies them to callback provided to it.
2. output_consumer(output) this function sends the output of the image processing to it's next destination

Notes

GRIP and SCVF compatibility

if you are using grip to generate your pipelines keep in mind that the pipelines generated from it aren't compatible with scvf out of the box due to differences between python3 and python2

as of now, these changes are required:

• change the enum for the blur type to a python3-enum
• change the findContours function as instructed here
• add a get_output() method that is compatible with the API specified above

Opencv 4

• Switch between the im2 and contour contours variables in the filter_contours function of your pipeline

IO implementations

• make sure that your custom output_consumer is compatible with the output provided by your pipelines.