Spherical to simple side images

Bunch of scripts to process and convert Mapillary spherical images into cube imagen and then into simple images for specific area in Mapillary API.

Build container

export MAPILLARY_ACCESS_TOKEN="MLY|..."
docker-compose build

Download points from Mapillary into geojson file

docker run -v $PWD:/mnt/ -e MAPILLARY_ACCESS_TOKEN=$MAPILLARY_ACCESS_TOKEN -it developmentseed/spherical2images:v1 \
    get_mapillary_points \
        --output_file_point=data/Warrendale_points.geojson \
        --output_file_sequence=data/Warrendale_sequences.geojson \
        --bbox=-83.2469680005052,42.3289420003625,-83.2157740004676,42.3578449996934

Simplify density of points

docker run -v $PWD:/mnt/ -it developmentseed/spherical2images:v1 \
    simplify_points \
        --input_points=data/Warrendale_points_filter.geojson \
        --output_points=data/Warrendale_simplify.geojson

Clip Pano images into left an right side

docker run -v $PWD:/mnt/ -e MAPILLARY_ACCESS_TOKEN=$MAPILLARY_ACCESS_TOKEN -it developmentseed/spherical2images:v1 \
    clip_mapillary_pano \
        --input_file_points=data/Warrendale_points_filter.geojson \
        --image_clip_size=1024 \
        --output_file_points=data/Warrendale_points_fixed.geojson \
        --output_images_path=data/ \
        --cube_sides=right,left

Note For converting spherical-pano image to cubemap image, we are using py360convert:

Example:

• Mapillary spherical-pano image:

• Cube:

Front	Back

Left	Right

Under	Over