Enterprise Europe Network

Monitoring system with intelligent machine vision detection and web-based interactive platform for analysis of waste sorting process workflow

Country of origin:
Country: 
SLOVENIA
Opportunity:
External Id: 
TOSI20201214001
Published
16/12/2020
Last update
11/01/2021
Expiration date
12/01/2022

Keywords

Partner keyword: 
Artificial Intelligence (AI)
Data Processing / Data Interchange, Middleware
Imaging, Image Processing, Pattern Recognition
Cloud Technologies
Internet of Things
Artificial intelligence related software
Process control equipment and systems
Machine vision software and systems
Chemical and solid material recycling
Process control and logistics
Other research and experimental development on natural sciences and engineering
EXPRESS YOUR INTEREST

Summary

Summary: 
A Slovenian R&D company from Primorska region specialized in process analytics in environmental industries developed a novel monitoring system with multiple detection devices (AI/ML machine vision) feeding a database with process information like belt load, waste composition, moisture and presence of contaminants. The solution is accompanied by a state-of-the-art Industry 4.0 information system. They are looking for a strategic partner for a commercial agreement with technical assistance.

Description

Description: 

To achieve the ambitious goals for recycling rates set by EU strategic documents, the processing of non-hazardous solid waste needs to be significantly improved. While majority of process industries are successfully implementing different digitalization solutions in order to optimize the processes in terms of efficiency and quality, these present a great and unused potential of improvement for waste processing. Real-time material flow monitoring for decision-informing is very rare, keeping the process fairly rigid and sub-optimal both in horizontal (across the entire value chain) and vertical (within facility) directions.
The current state-of-the art in process automation and control in waste processing is represented by PLC (Programmable Logic Controller) and SCADA systems (Supervisory Control and Data Acquisition), which are powerful tools for automation, but lack in ability to easily provide informative and advanced analytics. Additional information on the waste flow is available directly on sensor-based sorters, but this is limited to a single point in the process. In light of Industry 4.0 paradigm, a further step in data gathering and utilization will maximize efficiency of current technological processes, give insight into major bottlenecks and give rise to new business models stemming from the data utilization.
A Slovenian high-tech start-up with expertise in optical system design, AI (Artificial Intelligence) algorithms and data analysis, developed a complete monitoring system for waste sorting process with configurable detectors and data analysis platform.
The detection system currently consists of visible light and a 3D camera and can be extended upon request by a NIR (Near-Infrared) multispectral camera. Data-fusion of several detectors allows improved performance of AI algorithms for material composition.
Detectors and control electronics are fitted in a 400x500x180 mm box that is mounted on top of conveyor belt. A custom, energy-efficient LED illumination is supplied according to belt width and speed at the monitoring locations.
Features like belt load, waste composition, moisture and presence of contaminants are extracted in real-time and populate the data-base. Relevant production performance metrics are calculated using recorded process parameters and additional measurements.
Selected measurements and analysis results are displayed in an interactive web-based platform, where alarms are set for the case of abnormal process status. Secured login allows different user-profiles to access various real-time information via the dashboard or historical data via pre-configured or customized reports.
The most common usage cases of the monitoring system are assessment of process performance upon varying feeding material, optimization of process parameters with real-time feedback to derive different process recipes and quality assurance of sorted material to reach buyers requirements.
The system reveals previously in- or hardly accessible process information in real-time, while historical data are easily available for detailed analysis. Process optimization, dynamic maintenance scheduling and easier return-of-investment assessment that such data allow will bring immediate benefits like increased throughput, reduced down-time and minimized risk.
The company is looking for a strategic partner for a comm. agreement with technical assistance. The complete information (data monitoring, processing and interface) system will be installed in a partner’s waste sorting facility at optimal locations to extract partner’s information of business significance defined at the beginning of cooperation. The cooperation will focus on business objectives and production goals which will be directed to maximizing profit and reducing cost and risk. Partner’s active involvement with internal process know-how is crucial for definition of the abovementioned metrics. A case-study (whitepaper) with process performance analysis will be pu

Advantages & innovations

Cooperation plus value: 
The information system, based around our in-house developed middleware, allows for monitoring, aggregation and optimization of user-specifc measurements inside a waste sorting facility. It is of modular structure. If existing data is not available, the facility can be equiped with several low-cost multi-detector devices, placed on specific locations. The locations are selected optimally for the problem (s)he would like to solve. Such multi-detector configuration makes for low-cost monitoring of complex process scenarios. An example of a multi-detector device is a sensor-fusion of visible and hyperspectral cameras, providing information superior to human inspection. AI alghorithms trained with fused sensor data can outperform other single-detector systems. If existing data is available (from NIR sorters, robots, belt scales, etc.), they can be integrated into the information system and thus connected with other process parameters. Most of the currently available systems offer device-based monitoring and are not designed for complex business-centric measurements and optimisation analytics, thus presenting a barrier to the decision-makers. By using data fusion and advanced data science the system is able to provide complex insights into the process and thus directly enable decision-makers with solid information on-demand. By using data science and big data approaches in technological process monitoring, the users are expected to enjoy the following advantages: - perform data-backed upgrades, maintenance and planning - gain deep insight across their whole technological process and thus enable experts to focus on maximizing throughput of the existing sorting lines - generate new business models based on the solid and quantitative information provided by the system

Stage of development

Cooperation stage dev stage: 
Prototype available for demonstration

Partner sought

Cooperation area: 
The Slovenian company is looking for a strategic partner among medium sized and big (multinational) companies with one or more waste sorting facilities for plastic packaging, glass or paper, where monitoring system will be installed. Tight collaboration with partner is expected in the development phase of specific KPIs (Key Performance Indicators) and in support of performing field-measurements (waste classification by physical characteristics). Partner is supposed to provide all relevant information and access to other machine-data.

Type and size

Cooperation task: 
>500 MNE,251-500,SME 51-250,>500