LAPIN

What it is

A framework for the analysis of on street parking occupancy via Licence Plate Recognition (LPR) data.

Getting started

Requirements

Install Lapin python's requirments with conda :

conda env create --name <YOUR_ENV_NAME> -f environment.yml

Optional

You may need to have a docker installation available on your machine. See Valhalla mapmatching.

Configuration

Project configuration

Create a config file for your project . You can create a blank one by running :

python -m lapin -c

Mapmatching configuration

You have the choice between two mapmatching engine : OSRM and Valhalla. The main difference being that while using Valhalla you can do the matching directly on the Montreal Geobase. Doing so improve the accuracy of the positionning of the plate on the geobase. Thus improving the quality of the results.

Valhalla

To use valhalla, you'll need to compute the OSM network from the geobase file. Then create the valhalla graph with valhalla engine and the OSM network. The step are the following :

1. Create the OSM graph

python -m lapin --generate-graph

2. Generate Valhalla's graph

sudo docker run --rm --name valhalla_gis-ops -p 8002:8002 -v $PWD/data/network/valhalla:/custom_files -e tile ghcr.io/gis-ops/docker-valhalla/valhalla:latest'

3. Specify the use of valhalla in lapin/__main__.py line 113-115.

    matcher_host='<PATH_TO_LAPIN>/lapin/data/network/valhalla/valhalla_tiles.tar',
    matcher_client='valhalla',
    matcher_kwargs={'service_limits':{"trace": {"max_shape": 26000}}}, # your desired config

OSRM

To use OSRM simply identify a valid OSRM instance.

1. Specify the use of OSRM in lapin/__main__.py line 113-115.

    matcher_host=<ADRESS_TO_OSRM_INSTANCE>,
    matcher_client='osrm',
    matcher_kwargs={},

Note : the instance must be launched with a sufficiently large max-matching-size parameter (e.g. 100000)

Lauching an analysis

Then excecute the package with the following command :

python -m lapin --conf-file <PATH_TO_YOUR_CONF_FILE>

Installing the module

Clone the repo and install the lapin package.

cd <repo_dir>
pip install .