Skip to main content
European Digital Innovation Hubs Network

Harvesting Success: Piedmont's Digital Agriculture Breakthrough

Harvesting Success: Piedmont's Digital Agriculture Breakthrough
Published at 04 April 2024 | Italy

General details

EDIHs involved

Customer

Company name in green letters, a tree on top of the name
Customer type: PSO
Customer size: Medium (50-249)
Customer turnover: Not applicable, No sales – 600m€/year delivery benefit 

Services provided
Test before invest
Technologies
Internet of Things
data
Sectors
Agricultural biotechnology and food biotechnology

Challenges

Agenzia Regionale Piemontese per le Erogazioni in Agricoltura is the Piedmontese Regional Agency for Agricultural Grants. The Agency aims to support the agricultural sector (it delivers approximately €600,000,000 a year in benefits), complying with regional, national and EU regulations on funding. CHEDIH focuses on the agriculture, food and health sectors. The ‘test before invest’ service catalogue includes advanced digital technologies designed for these sectors. ARPEA needs to verify that certain plots of land are effectively employed for their intended use, for example, for the cultivation of beans. Since 2019, EU policy requires an ‘Agriculture Registry’. The EU Regulation of 2 December 2021, n. 2116, Section III, Art. 70 refers to a “system of surface/land monitoring” and establishes the framework and quality requirements of the monitoring system. There are new requirements arising from the new Common Agricultural Policy (CAP) for 2023-2027. They introduce innovations to encourage the adoption of faster and more effective procedures, and tighter deadlines for funding requests. Member States are also granted greater autonomy to simplify and rationalise the management of funds, to ensure compliance with EU rules.

A single implementation tool is therefore needed that is aligned with the digitalisation strategy in the Rural Development Complement (CSR) for the Piedmont Region (Art. 114, Letter b).ARPEA monitoring has been done using graphical representations of the territory (pictures) to interpret the use of land, followed by on-site checks. This poses two main challenges: not all areas are covered and there are limitations to surveillance and control activities conducted using satellites and continuous remote sensing. The resolution of the satellite images is not good enough to verify the dynamic use of land and is therefore ineffective and inappropriate for small lots and hills. A tool is needed to automate and accelerate required  eligibility checks.

Solutions

The technology adopted is defined as a ”satellite magnifying glass”. Satellites orbit at hundreds of kilometres from the Earth’s surface, as part of scientific and continued missions (typically Copernicus Sentinel 2). They are able to provide multispectral monoscopic images with a geometric resolution (GSD, Ground Sampling Distance) of up to 10m. However, the task of describing the average behaviour of crops with a different level of accuracy depends on the crop itself. Crops such as maize, wheat, barley, soybean, etc., can serve as a proper benchmark for satellite data. In contrast, the satellite segment does not allow for a precise investigation of crops such as orchards and vineyards. Being within a field pattern where fruit trees/vines alternate to an interleaved background, they can manifest as noise during signal interpretation. In these cases, a higher resolution (GSD of fewer centimetres) coupled with the stereoscopic capacity of airborne cameras, including both RGB (Red, Green, Blue) and multispectral, can offer a more accurate and efficient alternative. The airborne survey results are a highly effective solution because image interpretation (crop classification) can be performed automatically, using AI-based approaches to analyse high resolution data.Controlling CAP-related Geo-Spatial Aid Applications (GSAA) from farmers is a significant part of ARPEA’s work. The use of spotted and distributed high-resolution images from aerial cameras could greatly support, and possibly even substitute, expensive field campaigns needed to collect data to support satellite-based deductions (training/validation set). Within a single flight, it is then possible to map thousands of hectares (eventually non-contiguous) utilising General Aviation aircraft equipped with the same sensor-based technologies that are carried aboard satellites. This technology enables high resolution mapping of remote parts of the territory with a high level of detail. 

Results and Benefits

Overall, the adoption of the solution tested enables ARPEA to evaluate mapped territories more effectively, in terms of:

  • Integration of monitoring results (AMS: Area Monitoring System) with those of the remote sensing service provided by the current experimentation and the OTSC (On The Spot Check) to strengthen the AMS (monitoring system) and the training of classifier codes for land use identification.  This will allow the verification of land use on private lots prior to calibration on-site, reducing the number of more costly on-site checks;

  • Reduction of anomalous results resulting from AMS;

  • Positive impact on LPIS (Land Parcel Identification System), AMS (Area Monitoring System), and QA (Quality Assessment);

  • Reconstruction, via digital elaboration, of the third vertical dimension to enable the recognition of specific crops. ARPEA requires this data to pay coupled aids and to identify buffer zones;

  • Identification of water borders and maintenance (thanks to the third vertical dimension), as required by law, by measuring altimetric levels along the river axis to verify riverbank slopes;

  • More precise metric measurements. The results of the checks are returned through a GIS layer indicating the outcome for each plot.

Perceived social/economic impact

The adoption of the technology by ARPEA has several socio-economics impacts. It is a real boost to the digital maturity of ARPEA and can also be a channel for reaching out to farmers who are looking at ARPEA as an inspiration to kickstart their digitalisation. The benefit will be seen in a period of 1-2 years, prompting a DMA in years 2 and 3. 

The benefit will not only be for the work of ARPEA, which distributes subsidies to farmers, but also for the territory and the entire supply chain (CAA, funding agencies, Piedmont Region, Certifiers). The digitalization of monitoring activities will not only uncover fraudulent declarations about land use, but it will also provide an incentive to farmers to stick to the rules, resulting in cost savings for the public fund. This will result in big savings in terms of EU funds.

Furthermore, ARPEA was able to expedite and improve the quality of verifications without relying on on-site visits for land use verification. Over time, the data could be made available to the territory for further data analysis, and artificial intelligence techniques could be applied. For instance, to monitor crop health, enhance irrigation water usage, in a precision agriculture approach. The overall end result will be a more accurate calculation of payments due to ARPEA beneficiaries, producing cost savings due to the avoidance of fraud.

Measurable data

The outcome of the technology adoption has been estimated at the start. Measurement is in progress. 

The study has been classified as a high priority, due to the benefits it is generating: the identification of the real position of the riverbank edge and construction of the corresponding linear geographical layer; subsequent definition of the 5-meter buffer on which the commitment to maintain the buffer strip and the prohibition of treatments (BCAA4) is applied. This data is currently not available, making impossible to precisely identify plots subject to the obligation.

Output: High resolution (centimetre) true orthoimages coupled with a DSM from aerial stereoscopic images are very new sources of information that are supporting several tasks that ARPEA was not previously able to perform

1) Survey of sample fields to support satellite-based deductions about GSAA.

2) Detection and characterization of orchards/vineyards.

3) Semi-automatic detection of specific patterns related to orchard/vineyard traits/features. 

4) Detection and vectorization achieved using image segmentation or AI-approaches. Some immediate statistics concerning polygon patterns (tree crowns) can be easily extracted using ordinary GIS tools. 

5) Detection and mapping of the actual (and precise) position of the riverbank edge (understood as "the point of the bank of the incised or active riverbed at a higher altitude"). The BCAA4 conditionality defines a 5-metre-wide buffer around the riverbank edge where farmers cannot provide treatments. This type of accurate positioning could help farmers to comply with both BCAA4 and ARPEA controls. 

Sufficiently accurate information is not currently available. Indeed, the adopted reference corresponds to the hydrographic network layer of regional technical cartography, which is not consistent with the buffer width. The pictures below show an example based on the adoption of an interactive tool (profile extractor) that combines the image content from the RGB orthomosaic with the height information from the associated DSM to analyse local topography of riverbanks. 


6) Detection and mapping of erosion effects over field. The high level of accuracy and resolution provided by aerial images makes it possible to analyse borders and look for predictors of significant erosion from rainfall, including in summer time when crops are present in the field. A significant portion of any type of field, around the seeded area, can now be recognized and qualified from an erosion point of view, by looking jointly at the orthoimage and DSM.   
7) Qualification of pastures, in terms of the presence of trees (net/gross pasture areas).



The above images illustrate a case analysis from Very High Resolution (VHR) photos, without the third-dimensional data. It is unclear whether the positioning of the riverbank edge should be attributed to the blue or green polyline. Subsequently, the DSM image derived from aerial flights is presented (processed with shading), which will allow the maximum height data to be obtained, and will enable automated soil classification to resolve doubtful cases from satellite monitoring. 
Determining land use through remote sensing techniques based on free satellite data from the Sentinel 2 mission of the EU Copernicus Program still exhibits inadequate accuracy for certain crops with similar phenological developments. Basic climatic trends have shown moderate variability.

DMA score and results - Stage 0

DMA at start T0 : 48% score. 
Lowest score : Interoperability : 28%. 
Highest score : Data management and security : 78%.
 

DMA score and results – Stage 1

Not yet available.

Lessons learned

DOs
It was useful to set up a small advisory team of partners to evaluate whether the customer’s needs can be appropriately addressed by and whether that fits within the EDIH’s mission. Two  CHEDIH partners evaluated the eligibility of ARPEA to get access to the Digital Maturity Assessment tool (this was carried out via a profile interview with predefined criteria for selection).

The DMA tool helped to encourage ARPEA to set up a multi-disciplinary team of experts and ensure the maturity assessment was both representative and accurate. They had never had a DMA done before.

After the DMA, a description of their needs was requested using a predefined ‘template’. CHEDIH’s catalogue was then analysed thoroughly, in collaboration with the customer, to identify ‘matching’ technology for the ‘test before invest’ step. 

It is recommended that predefined templates should be developed and used for the stages of customer onboarding.

It was important for CHEDIH to embrace a ‘partnership spirit’ to ensure effective cooperation between the DMA provider and the project coordinator, especially at the early stage of the project (May 2023) when guidelines were not yet published by the local Ministry, in order to proceed with documentation and monitoring.

CHEDIH plans to provide ARPEA with training services, as well as support to find funding, by involving ARPEA in project proposals for ERDF funding managed by Piemonte Region.

DO NOTs

The EDIH customer must be guided through the onboarding process and through the requirements identification and service selection processes. Simply letting prospective customers choose services from the EDIH's catalogue leads to suboptimal outcomes and does not exploit the philosophy behind the EDIH programme.

Other Information

The customer’s interest in CHEDIH’s ‘test before invest’ service was facilitated by the dissemination activities carried out by two CHEDIH partners:

  • MIAC – a regional innovation cluster that represents farmers and the agrifood industry community in the region, and that understands the challenges associated with the optimal use of land for cultivation;

  • CSI – the regional in-house ICT provider that conducted ARPEA’s DMA and understands the digital needs and challenges associated with the adoption of new technologies by public administrations.

Given the public nature of ARPEA, CHEDIH will help the customer to find public investments to fund the tested solution. That will include informing ARPEA about funding opportunities, as well as helping the customer to form partnerships and participate in project proposals for EU or national and regional calls. The test was carried out in the 5 areas selected to be representative of a variety of Piedmont. 

Need support?

Consult our catalogue to locate the Eupopean Digital Innovation Hub nearest to you and accelerate your company's digital transformation.

Find my nearest EDIH