This Case Shows Specific Expertise. Find the Companies with the Skills Your Project Demands!

You're viewing one of tens of thousands of real cases compiled on Many.dev. Each case demonstrates specific, tangible expertise.

But how do you find the company that possesses the exact skills and experience needed for your project? Forget generic filters!

Our unique AI system allows you to describe your project in your own words and instantly get a list of companies that have already successfully applied that precise expertise in similar projects.

Create a free account to unlock powerful AI-powered search and connect with companies whose expertise directly matches your project's requirements.

AI-Powered Plankton Detection and Classification System for Underwater Monitoring

itransition.com

Environmental Services

Energy & natural resources

Government

Challenge

OceanSight Systems' existing plankton detection and classification software, based on older CNN algorithms and limited by hardware constraints, results in unsatisfactory accuracy and processing speed. The current system's 810 FPS image processing speed, limited by the C++ software and the camera's 30 FPS maximum, hinders timely and reliable data analysis crucial for ecosystem monitoring and informed decision-making.

About the Client

A US-based company specializing in the design and manufacturing of advanced underwater measurement and observation systems for real-time data acquisition in challenging ocean environments. Their systems are used by scientists and companies to understand ecosystem dependencies.

Objectives

Develop a robust and accurate AI-powered plankton detection and classification system.
Improve image processing speed to match the camera's 30 FPS capability.
Maximize the utilization of the NVIDIA Jetson hardware resources (CPU and GPU) for optimal performance.
Enhance the system's compatibility with the existing underwater measurement and observation platform.
Deliver a scalable and maintainable solution for future updates and enhancements.

Functional Requirements

Real-time plankton detection and classification.
Accurate identification of different plankton species.
Quantification of plankton density.
Data visualization and reporting.
Integration with OceanSight's existing data acquisition system.
Ability to save detected plankton images and statistical data.
User-friendly interface for system monitoring and control.

Preferred Technologies

Darknet (for plankton detection)

PyTorch (for plankton classification)

Faiss (for nearest neighbor search)

NVIDIA Jetson platform

PyCapture (for camera interface)

Integrations Required

OceanSight's existing data acquisition system
Potentially a cloud-based data storage and analysis platform

Non-Functional Requirements

High accuracy (defined as a minimum of 90% for plankton classification).
Real-time performance (processing speed should closely match the camera's 30 FPS).
Scalability to handle increasing data volumes and new plankton species.
Robustness and reliability in harsh underwater environments.
Secure data transmission and storage.

Expected Business Impact

Successful implementation of this project will significantly improve OceanSight Systems' ability to provide valuable insights into aquatic ecosystems. The improved accuracy and speed of plankton analysis will lead to more timely and reliable data, enabling better informed decision-making for clients in various industries, including environmental monitoring, resource management, and scientific research. This will enhance OceanSight’s competitive advantage and strengthen its position as a leading provider of underwater observation systems.