Accurately assess agriculture risk

Intro

Develop crop insurance products with UP42

In the face of increased climate variation and extreme weather, industrial agriculture companies are neutralizing risk through crop insurance. The pricing of crop insurance policies can be remotely informed by historical analysis and predictions of crop health, weather events, and water-related geohazards, such as flooding or irrigation contamination.

Overview

Access years of high-resolution archive imagery

Leverage the breadth and depth of UP42's archive data library, with satellite imagery, weather data, aerial data, and more - some of which have data going back over a decade.

Task satellites for imagery to assess insurance claims

React to insurance claims remotely by tasking high-resolution satellites for specific areas of interest, ensuring low cloud cover and specific incidence angle ranges.

Automate analysis with machine learning

Process your data with out-of-the-box machine-learning algorithms focused on crop health, geohazard predictions, and more.

Benefits

Benefits for you

Offer your customers a holistic view of their area of interest by not limiting the analysis to a single data set.
Use a wide range of out-of-the-box machine-learning algorithms to assess risk and predict damages.
Integrate outputs of your analysis into custom tools or your customers' claims management software via API.
Ensure you have unobstructed imagery for the time you need through our bespoke tasking service.

Benefits for your customers

Get all the geospatial inputs required to assess an insurance claim in a single tool.
Assess risk and claims with confidence, knowing the insights are built on industry-leading data sources and analytics.
Gain access to the raw imagery as well as the outputs of models to validate and justify findings.

Build your solution with UP42

The technical details

How can you use UP42 to build crop insurance solutions?

Data

High-resolution optical and radar imagery

Access SAR data from Sentinel-1 and optical data from Sentinel-2 to power different machine-learning models. Augment the machine-learning models with high-resolution Pleiades imagery, for which an extensive archive of imagery can be called upon for historical analysis. At the same time, you can also commission imagery for a given future period.

Historical analysis

Analyze historical crop health

Apply NDVI or other band math processing blocks on your historical Pleiades imagery captured on multiple dates to build up a view of historical crop health and ascertain the likelihood for that level of crop health to continue in the absence of geohazards or extreme weather variability. Additionally, use flood mapping to determine the historical prevalence of flooding using Sentinel-1 SAR images over time.

Forecasting risk

Machine-learning-backed risk assessments

Predict geohazards month in advance using Platoi's Advanced Water-Related Geohazards Predictor. This machine-learning algorithm leverages Sentinel-2 data to predict instances of flooding or exposure to water bodies. This can be a crucial input into a decision whether or not to approve a crop insurance policy.

API

Automate, embed and scale your analytics

Through the UP42, and supported by the Python SDK, it is easier than ever to integrate the outputs of UP42 workflows into your end-users' products. Trigger your analysis through the Python SDK and access your outputs via download or integrated into your products, cloud storage, or legacy insurance claims management software.

Access our Python SDK

import up42
up42.authenticate(project_id="12345", project_api_key="12345")
project=up42.initialize_project()

workflow = project.create_workflow(name="Crop Insurance", use_existing=True)
input_tasks=['oneatlas-pleiades-fullscene','pansharpen', 'ndvi']
workflow.add_workflow_tasks(input_tasks=input_tasks)
parameter = workflow.construct_parameters(
    geometry=[13.4466, 52.495598, 13.453016, 52.499216],
    geometry_operation="bbox",
    scene_ids="DS_PHR1A_202005061018418_FR1_PX_E013N52_0513_01179")

job = workflow.run_job(input_parameters=parameter, track_status=True)
job.download_results()