How can Satelytics be so accurate? This is the most common question we get during conversations with prospective clients, usually asked in a skeptical tone. Satelytics was founded on the belief that accuracy is the priority in every project we undertake. Here is a glimpse into what makes for accurate algorithms.

Our Constant Vigilance™ algorithms provide alerts to physical changes, methane, liquid leaks, vegetation, nutrient loading, metals, and more, all by analyzing commercially available satellite imagery. The phrase “commercially available” is key. We source imagery from any number of satellite data providers depending on the challenge the customer is trying to address. Remaining agnostic to the imagery source allows Satelytics to be adaptable to each customer’s needs.

Before any of our algorithms are offered commercially, a strict development program rooted in “ground-truthed” (calibrated) measurements is completed. At the heart of our algorithms is a convolutional neural network that requires training data for calibration. To gather this data, our scientists first source publicly available spectral data sets. We then complete fieldwork to gather and analyze physical samples to be analyzed independently by a commercial laboratory.

On location, our scientists gather spectral samples with a backpack spectrometer while simultaneously tasking satellite imagery over the AOI to calibrate data points. In many instances, we task from several vendors simultaneously to offer surveying flexibility in the future.

Attention to detail like this comes at a cost when developing algorithms. Satelytics will never claim to be the least expensive geospatial analytics solution available. We work hard to ensure that what it says on the outside of the box is matched or exceeded by what it does on the inside.

Give us a call today to learn about many real-world examples proving the statements made here. We’ll walk you through the industry-proven results.