Intersection over Union for Image-annotated crowns

[bw4sz]

We will try to keep the inputs and outputs pretty similiar to R package to help with debug.

> library(NeonTreeEvaluation)
> head(submission)
       xmin     ymin     xmax     ymax     score label plot_name
1 217.24730 265.0254 290.4331 337.5056 0.7930149  Tree  ABBY_020
2 284.37094 339.3117 338.5593 388.7977 0.7155186  Tree  ABBY_020
3 197.60432 350.3243 249.3276 400.0000 0.7128302  Tree  ABBY_020
4  25.21722 186.6763 123.3826 283.8552 0.6598154  Tree  ABBY_020
5 332.43198 293.3734 377.3869 343.7137 0.6279798  Tree  ABBY_020
6 198.98254 181.8541 266.0995 249.9332 0.5793932  Tree  ABBY_020

#Compute total recall and precision for the overlap data
results<-evaluate_image_crowns(submission = df,project = T, show=F, summarize = T)

doc is here

https://github.com/weecology/NeonTreeEvaluation_package/blob/5f36acfe170aaec04cee1800a64cfb615b7fd52e/R/evaluate_image_crowns.R#L2

[bw4sz]

label is actually a placeholder for some future where we do species class. Its not required in any code.