INSPIRE Environmental (INSPIRE), a Venterra Group company and team of marine scientists focused on seafloor health, used Microsoft Azure solutions to streamline the analysis of complex ocean data. INSPIRE, which collects large volumes of data to help energy companies and government agencies understand the potential for their projects to impact seafloor health, was increasingly challenged in processing and analyzing the data. With exponential growth in offshore wind and other projects affecting undersea environments, INSPIRE created a solution allowing it to analyze and interpret seafloor images more quickly and efficiently. INSPIRE used AI for image analysis and improved business by lowering costs and reducing turnaround times for clients.
Assessing the benthos, the environment of the seafloor, is crucial for promoting environmentally safe energy options and using sea-based resources responsibly. INSPIRE Environmental (INSPIRE) has deep expertise in understanding the complex ecology of the seafloor, providing data analysis and accurate visualizations of benthic habitats to private industry and government agencies.
The data INSPIRE provides is essential for assessing seafloor health, but collecting and analyzing it is time-consuming. INSPIRE wanted to make the collection and analysis more efficient while providing its clients with faster results, so it sought ways to automate labor-intensive tasks.
By developing a solution on Microsoft Azure, INSPIRE significantly reduced costs, processing times, and labor.
Seeking new ways to analyze and store tens of thousands of seafloor images
INSPIRE works with five main ocean economy sectors: offshore wind, offshore oil and gas, dredged material management, seafood waste, and contaminated sediments. It supports clients in conducting baseline assessments, post-construction monitoring, and similar tasks that enable the completion of wind farms and other energy development projects in ways that protect the seafloor. “We conduct baseline studies for new projects and help parties better understand the potential impact of the development on the ecosystem,” says Jeanine Boyle, CEO of INSPIRE.
To collect data, INSPIRE uses Sediment Profile Imaging, an innovative technology developed by one of its founders and now an international standard for seafloor assessments. “We image an area of the seafloor with a downward-looking camera that we then plunge into the seafloor to collect an image within the sediment,” says Kersey Sturdivant, Principal Scientist at INSPIRE. In each survey, the team collects 1,000 to 2,000 images. “The scale and scope of the data we collect is massive,” says Sturdivant. “We regularly collect terabytes worth of data in a single survey.”
INSPIRE assesses how sea life functions within the soft-bottom seafloor. However, assessing data can be challenging. “A person typically sits and annotates on a screen,” says Sturdivant. “It’s very labor-intensive and time-consuming.” Faced with this challenge, INSPIRE was determined to find a way to automate those tasks and create a more efficient data-analysis process.
Using Azure AI to automate data analysis
INSPIRE has a close and long-standing relationship with Infused Innovations (Infused), a strategic advisory company and Microsoft partner. In 2019, Infused met with INSPIRE to discuss whether the company was analyzing data in the most efficient, cost-effective way and whether it could use Microsoft technologies to create a better solution. With support from the Infused and Microsoft teams, INSPIRE conducted an initial proof of concept (PoC) in early 2020 that showed promising results.
In the summer of 2022, INSPIRE decided to pursue Azure AI. Increasing demand for offshore wind power was rapidly expanding its client base, and INSPIRE believed that AI technology had matured to a point where it could serve its needs when used with its own tools.
INSPIRE and Infused created a minimum viable product. “Azure services are incredible,” says Jeffrey Wilhelm, Founder and CEO of Infused. “INSPIRE already had great data and tools, and by deploying Azure, we streamlined its image analysis and processing workflow to build the solution it had dreamed about.” The quantity of images that INSPIRE had collected and its effective data-management practices were key to its ability to use Azure AI.
Deploying new solutions to exponentially scale analysis capability
INSPIRE set out to automate the process of annotating the images it uses to assess seafloor health. It wanted to make the task less time-consuming so that its analysts could concentrate on interpretation and INSPIRE could meet industry demands for faster results and real-time decision-making. “The desire was to add efficiency to the process and provide quicker results,” says Wilhelm.
After validating results through the PoC, INSPIRE deployed two case studies. The first used Azure Custom Vision, an AI tool based on Azure Cognitive Services. Azure Custom Vision is an image recognition service that helps users quickly build, deploy, and improve their own image identification models. INSPIRE used this technology to build models that could predict grain size.
The second case study used automated machine learning through Azure Machine Learningto identify sediment boundaries within images. The results from those case studies paved the way for an automated annotation and model creation process that allowed INSPIRE data scientists to focus more on rigorous quality control than on labor-intensive annotations.
Providing high-quality results more rapidly
The system INSPIRE developed using Microsoft solutions made maintaining the quality on which the company built its reputation less time-consuming. “We were more efficient on analyzing the images, so we could put more effort into ensuring data quality,” says Alex Rhoads, Data Specialist and Statistician at INSPIRE.
Using Azure Files, a serverless file-share system, INSPIRE easily stored the extensive collection of images used in building and training the AI model. Now, INSPIRE seamlessly updates seafloor data that’s accessible to clients in near real-time. “We’re constantly working to get data to our clients so they can make informed decisions,” says Sturdivant.
Deploying Microsoft services has also helped INSPIRE improve internal business practices. “These tools help us perform our analysis at a lower cost and grow by adding more projects per analyst,” says Boyle.
Expanding machine learning models and meeting increasing demands for visual data with Azure
Azure solutions are a big factor in INSPIRE’s strategic planning. “The demand for visual data is ever-increasing, as is the ability to process that data,” says Sturdivant. “We’re working to meet that demand with an understanding of where it’s headed in the future.” INSPIRE is now ready to face future demands with new and evolving technologies. “It’s worth the investment,” says Boyle.
INSPIRE intends to continue creating models for additional use cases and increase its capabilities in automated analysis. INSPIRE also hopes automation will pave the way for new and innovative approaches. “We’re going to have a much more powerful, flexible, and adaptive product,” says Sturdivant. “That will provide innate scalability or integration with other systems as we continue to grow and develop.”
“The desire was to add efficiency to the process and provide quicker results.”
Jeffrey Wilhelm: Infused Innovations
Founder and CEO