Wild Nature Analysis
This organization aims to enhance understanding of animal behavior in wild nature by collecting and analyzing as much visual data as possible. The primary purpose of this project is to capture the beauty of biodiversity and help to preserve endangered species. It collaborates and partners closely with Yale University, Google, museums of natural science and history, and other similar organizations.
Client location
USA
Industry
IoT
Duration
8 Month
Team
6 people
Challenge
There are multiple camera traps for wildlife capturing all over the world. They usually belong to local communities, natural reservation areas, researchers, scientists, etc. Each day, thousands of new photos are captured, but they are not effectively shared and analyzed. Therefore, an initiative was launched to extract more value out of the available content. They needed a solution that would enable people from anywhere to upload their photos of wildlife to a unified platform.
Solution
Four Ages was glad to help design a solution that could launch this socially responsible mission. We have taken care of the end-to-end project work, from requirements gathering to implementation and testing.
Cloud architecture
The platform is cloud-based, so our architect and DevOps designed its structure. They considered the average number of new photos uploaded to ensure the overall cloud architecture could handle the data loads. Our experts developed modular, independently deployable services, also known as microservices, that run on Docker and Kubernetes to grant platform scalability and flexibility.
Back-end development
Our team used Node.js and Nest.js to design the back-end infrastructure for the platform. We built a robust back-end capable of handling large amounts of unstructured data, such as high-resolution wildlife images, and efficiently processing them.
Front-end development
The visual appearance of the website was also designed and developed by Four Ages. Our engineers primarily used React.js for most of the front-end tasks.
AI research and development
To extract value from the camera trap content uploaded to the platform, AI and ML algorithms were implemented. Specifically, TensorFlow and Keras were used for machine learning and model training based on thousands of images. The OpenCV library helped derive information from images in real time after completing the learning and training processes. These technologies enabled the automatic identification of animal species almost instantly.
Impact
The digital platform implementation for this project facilitated the global exchange of wildlife photos. Once an image is uploaded to the system, AI automatically identifies it within seconds, significantly reducing the time researchers spend on manual analysis.
By aggregating camera trap images, the platform provides access to timely data for those who need to monitor wildlife effectively. It also fosters a community where anyone can explore data and make informed decisions based on it. This comprehensive solution supports wildlife preservation and population recovery efforts.
Responsibilities
Cloud architecture
Design architecture of microservices
Frontend development
Backend development
CI/CD setup
AI research and development
Technologies
Author:
Maria Roy
