Recognition axe placement ORACs

Municipality of The Hague

Digital scanning and categorisation of types of byplacements at Orac's.

Last change on 23rd of August 2024, at 16:11 (CET) | Publication Standard 1.0

• Publication category: Impactful algorithms
• Impact assessment: Field not filled in.
• Status: Out of use

General information

Theme

Public Order and Safety

Begin date

Field not filled in.

Contact information

[e-mail]

Responsible use

Goal and impact

Exploit the technical development on image recognition in the operational process by using parking enforcement's scan cars. Image recognition can be a good addition to the various ways of spotting side placements, as scan cars are already driving around the city. The aim is to strengthen the MOR chain and speed up follow-up. Reduction on repeat notifications, duplicate notifications, manual actions and the follow-up time of clearing is -in addition to service to citizens- a significant improvement in business operations for the chain partners. With the application of scanning technology, the aim is to additionally feed the notification chain with -extension of- existing resources. Waste clearance and enforcement can potentially be followed up even faster.

Byplacements are detected and cleared more quickly (clean city);

The notifier is informed faster (better service), if a report is also detected by scanning technology.

Considerations

There are several initiatives to combat litter nuisance. This project is an addition in the approach and not an alternative.

The council passed a motion on 20 February 2020 in which the council called on the college to conduct a study on the deployment of scanning cars (RIS 304747), this study showed the technology was available. The college also wanted to investigate whether the data obtained from the scanning technology could be integrated with reports from the reporting system (so that together they form one source of information for planning the operation on the street) and/or the data could be used for information-driven work in capacity planning and policy-based approach to byplacements. The study was conducted and found that linking data can be achieved.

Human intervention

A deskforce monitors the image recognition in particular to promote the self-learning effect of the applied artificial intelligence on automatic assignment to the correct waste category and follow-up information provision to the relevant chain partners. This process can be stopped at any time.

Risk management

There were risks regarding storage of images, these have not yet been identified. The algorithm has been stopped.

Legal basis

The Environmental Management Act. This Act came into force on 1 March 1993 and provides an important framework for environmental management in the Netherlands. It provides the legal basis for environmental regulation and policy, including the management and collection of waste by municipalities.

Operations

Data

Image captures of oracs.

Technical design

The scanning vehicles are provided with the exact locations (GPS coordinates) where the ORACs are located. This is used to scan and determine when overview photos should be taken. Nine photos are sent from each ORAC; 3 from the camera facing forward, 3 from the angled camera and 3 from the camera facing backwards. It does not matter whether the ORACs enter the field of view of the cameras to the right or left of the scanning car.

The analysis of the images is done centrally. It is determined for each ORAC whether it has been superimposed. If so, it is categorised according to categories determined by The Hague (bags, boxes, chemical, white/brown goods and other). An indication of the volume of the additional placement is also given.