Vision on earthtoolsmaker

Racing Models, Not Opinions: How We Ran Wildfire ML R&D for Pyronear

Fri, 12 Jun 2026 00:00:00 +0000

In 2024, we wrote about building an early forest fire detector with the NGO Pyronear: cameras on antenna towers, a YOLO object detector (a fast model that draws boxes around what it finds in a single image) running on a Raspberry Pi, and real fires detected from 35 kilometers away. This post is the next chapter — and it is less about a model than about a method.

Over the past months we helped Pyronear answer one question: can a model that watches smoke evolve over time cut false alarms without missing fires? The answer turned out to be yes — the winning model raises 4× fewer false alarms than the production baseline while catching slightly more fires. But the part worth writing about is how we got there: a literature survey distilled into a shortlist, a standardized experiment harness, and a leaderboard where five candidate models — including the current production system — raced on the same frozen test set.

Smoke Is a Behavior: Inside Pyronear's Temporal Wildfire Detection Model

Fri, 12 Jun 2026 00:00:00 +0000

In the previous post we told the story of how the new smoke verifier for Pyronear was selected: a literature survey, five candidate models, one leaderboard, one winner with 4× fewer false alarms. This post opens up the winner itself — the bbox-tube temporal model (bbox for the bounding boxes a detector draws around candidate smoke, tube for the way those boxes are linked across frames) now released as a versioned package — and walks through how it judges a sequence of camera frames, stage by stage, with figures computed from real wildfire sequences.

Identify individuals with Local Feature Matching

Mon, 09 Dec 2024 00:00:00 +0000

In this blog post, we will explore a powerful technique widely used for identifying individuals across various species. This method leverages unique physical markings that remain relatively stable throughout an organism’s lifetime, making it effective for species with distinct patterns. For instance, whale sharks, trout, turtles, and seals all possess unique spot or scale patterns that lend themselves well to this computer vision approach.

How to prepare data for IDentification?

Sun, 08 Dec 2024 00:00:00 +0000

In this blog post, we will explore essential preprocessing techniques for normalizing animal images, preparing them for individual identification.

We will focus on the initial stages of the machine learning pipelines developed for various projects, specifically bear and trout identification. In both cases, similar computer vision techniques and strategies were employed to successfully create robust identification systems.

Gallery / Overview of the ML pipeline to identify bears using their facial markings with Metric Learning

Tracking the Journey: How to Monitor Wild Salmon Migrations

Fri, 30 Aug 2024 00:00:00 +0000

In this blog post, we will dive into the successful development of a wild salmon monitoring system, currently deployed to 8 different sites in British Columbia.

We must take every step in our control now as climate-change related trends will make things more difficult for salmon populations in the years ahead. This demands urgency for Pacific salmon and for the 130+ species, including grizzlies, orcas and eagles, that depend on Pacific salmon.

Protecting the Forest: Building an early forest fire detector

Wed, 21 Aug 2024 00:00:00 +0000

In this blog post, we’ll delve into the development process of a cutting-edge early forest fire detection system, created in collaboration with the NGO Pyronear.

Our detectors communicate fire alerts to a database that is connected to a supervision platform for the fire department.

– Pyronear

Pyronear offers a holistic solution for managing fire risks. Central to its capabilities is an innovative early wildfire detection algorithm, seamlessly operated on a compact microcomputer. This core system is augmented by a network of high-resolution cameras strategically positioned at elevated vantage points, providing panoramic coverage of forested regions. Together, these components form a resilient and proactive strategy for wildfire prevention and management.

A guide to designing a bear face recognition system

Fri, 12 Apr 2024 00:00:00 +0000

In this blog post we’ll delve into the technical development of a bear face recognition system which is a critical component of the bear identification system developed in close collaboration with the BearID Project NGO.

Our research and software tool will provide a replicable technique and general approach that can be applied to other species beyond bears, which could aid conservation efforts worldwide.

– BearID Project

For a comprehensive understanding of this project, please click on the pipeline overview below:

A guide to designing a bear face segmentation system

Thu, 11 Apr 2024 00:00:00 +0000

In this blog post we’ll delve into the technical development of a bear face segmentation system which is a critical component of the bear identification system developed in close collaboration with the BearID Project NGO.

Our research and software tool will provide a replicable technique and general approach that can be applied to other species beyond bears, which could aid conservation efforts worldwide.

– BearID Project

For a comprehensive understanding of this project, please click on the pipeline overview below:

How to build a benthic coral reefs analyser

Wed, 10 Apr 2024 00:00:00 +0000

Welcome to our blog post where we’ll explore the development process of a benthic coral reef analyzer, created in partnership with ReefSupport. Our goal is simple: to improve the tools available for monitoring coral reefs and marine environments. Let’s dive into how we’re making this happen!

For a comprehensive understanding of this project, please click on the image below:

Leveraging computer vision for the segmentation of coral reefs in benthic imagery holds the potential to quantify the long-term growth or decline of coral cover within marine protected areas

How to build a real time bear detection system

Tue, 09 Apr 2024 00:00:00 +0000

In this blog post, we’ll delve into the successful development of a real-time bear detection system, achieved through collaboration with the NGO HackThePlanet. Our initiative aims to safeguard Romanian farms by deterring bear encroachments.

Implementing non-invasive methods to deter bears from approaching farms and livestock holds promise in fostering harmonious relations between humans and bears.

For a comprehensive understanding of this project, please click on the pipeline overview below:

Project Scope

The computer vision model responsible for detecting bears will operate on a low-power microcontroller (such as the Raspberry Pi5). This necessitates swift operation and minimal power consumption during inference. Ensuring timely detection of approaching bears is critical, as their entry onto farms can result in predation on livestock, such as pigs. Achieving an extremely high recall rate is imperative to prevent such incidents. Additionally, since brown bear activities can occur day or night, the system must remain operational around the clock.