Feature/gpu detections and config

[mkutu]

Did a couple major things:

1. Added shapefiles for both locations. The shapefiles include transformations that @skovsens added. Technically, the shapefiles for each location are exactly the same. I copied and placed them in their respective location folders instead of using just one file to keep the directory structure intact.
2. I made minor changes to Species_info.json
3. added optional detection csv concatenation to automatically merge @skovsens detection results. This is done in localize_plants and set in config (cfg.data.detection.concat_detections)
4. added gpu support in how docker run is being called either with --gpus all or without.
5. and flexible GCP file retrieval in auto_sfm.py to accomodate differences in file name.

[savadikarc]

@mkutu I'm wondering if we need to keep the GCP csv files and the shape files in the repo.

[mkutu]

@savadikarc i put those to keep a working copy somewhere. But can see how it's probably not best practice. I can easily place those in the VM instead and keep the repo clean. I'll make a new commit and push

[mkutu]

@savadikarc I've made the changes:

• removed shapefiles and csvs
• removed removed QGIS project file and old zip file
• added a .gitignore to semif-utils