Zoo Workflow Enhancements

[2320sharon]

There are few minor enhancements that the team would like to have made to the zoo workflow. There enhancements do not change the fundamentals of how the workflow works, but instead make the outputs easier to use.

• Smooth the shorelines outputted by the zoo workflow using code developed by @mlundine
• Add a new column to the extracted shoreline geodataframe that includes the model score outputted by the classifier that sorts the good/bad segmentations.

[2320sharon]

Clarification: the scores that should be saved in the extracted shoreline geodataframe would be the model scores outputted by the good/bad image classifier model. The reason we would want to use these scores and not the Kmeans classifier score is because only the segmentations with the "good" kmeans classifier scores are used to extract shorelines. The Kmeans classifer outputs 2 scores and each image falls into one of those categories. If this score were to be outputted in the final extracted shoreline geodataframe all the shorelines would have the same score.

We want the good/bad image classification score so that we can sort shorelines based on how "good" the image's quality is and use this data to optimize our models.