2nd BEACON Business Case
The 2nd BEACON Business Case represents Serbian Case conducted together with the Agricultural Insurance company Triglav Osiguranje from Serbia. Within the case, the full range of BEACON services will be tested, focusing on the insurance against “Hail, Fire, Lighting” as 90% of damages in Serbia are caused by the hail.
In Serbia, Agricultural Insurance scheme represents the combination of public-private scheme where only one company is the representative of public insurance sector – DUNAV OSIGURANJE. It is presented on voluntary basis and as a commercial activity. The main characteristic of this sector is relatively low market penetration. The private companies are dominantly involved in provision of AgI, offering their insurance products to both legal and individual entities. The company Triglav from Serbia represents an early adopter of the BEACON toolbox. Triglav Insurance has existed in Serbia for 40 years through gaining the trust of an increasing number of clients from year to year. Triglav osiguranje is a Serbian insurance company that belongs to the leader in insurance in the Adria region - the Triglav Group.
For an effective hail event coverage, an image acquisition strategy is followed in BEACON to ensure that the pre- and post-damage imagery are representative of the insured crop’s condition. |
A 6-day pre- and a 20-day post-event composite image is produced by a 36-day and a 45-day pre- and post-event multi-temporal image stack, respectively. |
A Difference Percentage Index is calculated with the Vegetation Index differencing technique, by the pre- and post- hazard satellite image. |
The result is then used to feed a Machine Learning model for hail quantitative damage estimation. |
The Machine Learning classifier is applied for damage spatial distribution and severity classification in the final image. |
The BEACON approach for hail damage assessment was tested against ground-truth data, yielding high accuracy results during the early, vegetative and anthesis stages of wheat, maize and soybean. |
Machine Learning approach