Solid waste accumulation and illegal dumping in and around urban areas remains a key challenge for Municipal authorities' waste management services, negatively affecting the environment and degrading the citizen's quality of life. Tasbin, is a garbage detection and mapping platform that leverages artificial intelligence to provide authorities with the next generation tools for tackling the problem
One Page Summary
Waste collection and disposal is one of the main challenges that municipal authorities are faced with, on a day to day basis. Uncontrolled disposal of household and construction solid waste in residential areas, industrial zones and rural areas is taking a toll on the environment and the citizen’s quality of life. Despite waste collection planning by Municipal Authorities waste accumulation remains a prevalent complaint of citizens. For example, 25% of the complaints received by the Citizen’s Bureau of the Municipality of Ypsonas (CY) are related to waste accumulation.
The timely garbage collection, resource allocation and waste management procedures remain high priorities in the agenda of many municipalities in Cyprus and the rest of the EU. Overfill of designated disposal points is common at densely populated areas. However, beyond residential areas, the problem of illegal garbage disposal at non-designated locations, presents authorities with an additional burden that cannot be tackled without incurring significant additional costs.
According to EU statistics (2014), Cypriots produced ~2m tonnes / year, of which 31% are construction and demolition waste, 48.9% from other activities (inc. industrial waste), and just 20.2% is household waste. Cypriots hold the first position in the improvement of waste generation excluding major mineral wastes, reducing this waste to the 1/3 in 2014, compared to 2004. This demonstrates that Cypriots are environmentally conscious and thus a sound targeted group to test new waste reduction technologies.
Tasbin is an innovative garbage detection and mapping technology that leverages state-of-the-art hardware and software techniques to provide municipalities with a dynamically updated “garbage heatmap”. This heatmap is crucial for identifying accumulation sites, for devising optimal waste collection routes, better planning of designated garbage disposal locations and tackling of disposal in less populated suburban and rural areas.
Tasbin is a technology composed of 3 main components: (a) end-user applications for video and image data collection, (b) algorithms for image data processing and analysis for garbage detection based on Artificial Intelligence, and (c) an online garbage mapping and visualization platform.
Tasbin relies on 3 data sources each utilizing a different actor: (a) crowd-sourced images and videos of waste piles from citizens using a free smartphone application, (b) video captured directly in moving vehicles (inc. municipal trucks), and (c) video captured by semi-autonomous drones flying over industrial and rural areas.
An open application programmable interface (API) will be provided for integration of additional data, such as sensor data collected from smart trash bins, that are becoming popular means of monitoring designated disposal points.
The innovation of Tasbin lies mainly on the use of advanced Deep Learning algorithms and Computer Vision to automatically detect garbage overfill and identify uncontrolled garbage deposits using images and video footage. To reduce network traffic while facilitating real-time data processing and high-precision detection, Tasbin’s AI model leverages on both Edge Intelligence (lightweight AI running on mobile devices) as well as sophisticated Deep Learning algorithms utilizing powerful servers. This design allows immediate video processing and transmission of low-accuracy, and size-reduced processed data to the Tasbin platform, i.e. suitable for end-user smartphones, while servers can be used to collect and analyze large datasets to produce highly accurate results.
Geolocation of detected garbage, based on GPS data, will be correlated and presented to operators through an intuitive heatmap visualization platform using popular city maps (e.g. Google Maps, OpenStreet). Heatmaps overlaying municipal areas with accumulated waste will be dynamically updated as they are detected. Statistics and historical data will help identifying trends in waste generation and an alert system will keep operators informed about garbage accumulation across the municipality.
Tasbin can be instrumental in decision making and area-specific policy making for waste management. Most existing smart city solutions rely mainly on sensor data (i.e. smart bins), omitting the trash that may be placed around or away of those bins. With the sophisticated analysis of image and video as well as the crowd-sourced collection of data, Tasbin introduces a new low-cost and easy to deploy technology, which can be easily combined with data from existing systems, such as smart trash bins. Thus, Tasbin is an attractive proposition for municipalities and private waste collection service providers.
A Proof of Concept (PoC) of the Tasbin technology has won the 1st prize of RiseTech Media / CyRIC - Smart City Challenge among 13 competing teams. Through the current project the team aims to extend the PoC into a Mininum Viable Product.