Add json output for the detection boxes

[snlongmore]

Input:
Big orthomosaic with TIFF format
Step 1: Upload the Orthomosaic using webapp.py
Step2: Slice the image into smaller TIFF file called Tile by calling the function "Sliceing",
Output 1: several TIF file called tile.
NOTE: Here the Tile will keep Georeferencing information in Metadata as it is a .TIF file
Steps 3: call the "inject_fake_detection" function, this will put bounding box "as a picture". on each Tile (function to be replaced by final YOLO model).
Output 2: Tiles with bounding boxes
Step 4: Each bounding box will have precise location on the tile and the unit is "pixel", it does not have GPS coordinates.
Step 5: Extract the location of the bounding box based on pixel
Step 6: Extract the corresponding GPS coordinate of the bounding box location and saved it in JSON file
Output 3: JSON file,
NOTE: Each Tile will contain none or one or more than one bounding boxes
Step 8: combine tile and bounding box and save it as shapefile
Output 4: several shapefile with bounding box.
Step 7: Reassembling the orthomosaic using the Stitch function

Final output:

• Be able to add bounding boxes on the input TIFF image.

[Tsito-RAKOTO]

First version of solution have been created and pull request was sent for review. Please share your feed back. and feel free to share your comments.