How we leverage Spectralytics for Faster and Traceable analysis for crop insurance
Farmers routinely suffer crop damages from the environment (wind/hail), vehicle traffic, and animal migration/movements. Damage compensation from insurance policies are critical to fairly compensating farmers for lost crop revenues as a result of these damages. Estimating the extents of damages and their causes is critical for fair compensation and identifying the right insurance party (e.g. states typically pay for animal damage). Overhead imagery is a great tool for assessing damage extents.
But some GIS tasks, while important and critical for decision-making, are tedious and time consuming. For field damages, problem-owners will often have to “guess-timate” area coverages with varying accuracy. Computers are able to speed this up dramatically, but the nuance of the problem is tricky to code a solution for – that’s where our machine learning comes in!
How this works: This field (above) exhibits impacts from wind, animals, and vehicles on healthy crops. It’s easy for your eye to see the difference, but this is not so easy for a computer – after all, these are all instances of the same thing: flattened crops!
To train our machine learning models, we use human expertise to “annotate” a handful of sample images extracted from our data. Once done, we put the computer through its paces to learn how to generate the same annotations itself. This is a simplification of the process, but it illustrates how accessible the development process is.
Once we are satisfied with the performance of our trained model, we are ready to run it at scale across the cloud, to process as much data as we can possibly throw at it.
The value add: We manage the development and deployment of Machine Learning pipelines to automate the work of your analysts for your analysts.
Building these solutions yourself requires a multi-disciplinary team of engineers, and this combination of risk/time/investment is often preventative.
This is where Spectralytics shines! We combine years of expertise across DevOps/ML/GIS/Drones/Data-Analysis to build you the solution to simplify and accelerate your workflow – generate value quicker with less effort.
The examples and advancements in this particular use-case for AI are the result of solid teamwork between the data collectors, problem-owners, and data analysts.
We are passionate in using Machine Learning and AI to make a positive difference. We want to work with you and we are excited to see how far we can take our technology – we’d love to hear from you!
Recent Comments