Aerial drones and robots can collect massive amounts of data, but that information is only useful if it can be analyzed and passed along to humans for effective decision making. Kespry Inc. today announced the availability of Kespry Perception Analytics. The company said its system is designed for industrial use cases requiring comprehensive analysis of complex visual data, including asset condition tracking and identifying anomalies.

Menlo Park, Calif.-based Kespry was founded in 2013. It offers automation and machine learning capabilities to help businesses organize photos, videos, and infrared footage to identify and manage issues with physical assets. The company said its capabilities are based on its experience in drone-based aerial intelligence serving more than 270 mining and aggregates companies across North America, Europe, and Australia, as well as some of the world’s largest insurers.

Kespry customers include Colas, Fluor, Grinnell Mutual, Lehigh Hanson/Heidelberg, Oldcastle, Shell, Titan America, XAP 360, and Zellstoff Celgar.

Kespry Perception Analytics designed to streamline inspections

Instead of manually reviewing multiple sources of visual and sensor-based data, Kespry Perception Analytics is designed to streamline the inspection process by creating a geotagged, historical repository for visual data. This enables teams to easily analyze multiple sources of data across assets, track trends over time, and perform proactive maintenance.

At the heart of Kespry Perception Analytics is a knowledge graph, said the company. It maps a company’s entire library of visual data, including media files and photogrammetric output, by types of physical assets, their specific geographic location, and the types and times of issues identified.

“Kespry Perception Analytics delivers unprecedented business insight and solves major problems for industrial companies that have struggled to get meaningful value from visual data in a timely manner,” stated George Mathew, CEO of Kespry. “It provides companies with a more complete view of the state of assets than just depending on telemetry data alone. It’s designed with a simple interface to help users intuitively navigate through complex analysis with ease.”

The platform provides a comprehensive toolset to ingest and index the data, according to Kespry, and it uses Microsoft’s Azure machine learning to generate insights on the data. Kespry Perception Analytics’ intuitive search and analytics capabilities enable reliability and maintenance teams to query data without any coding knowledge, claimed the company. The system offers interactive dashboards and data visualization tools to analyze the health of assets across an organization.

“Until recently, industrial companies have solely relied on telemetry and sensor-based data, in addition to manual visual inspections, when monitoring their assets,” Mathew told The Robot Report. “Kespry Perception Analytics [KPA] redefines how companies perform predictive maintenance by creating a holistic repository for visual data, such as digital photos, videos, thermal imaging, and ultrasound, allowing operations teams to better organize and analyze important components of their inspections.”

“KPA captures data from virtually any camera-based sensor, including cameras based on drones, as well as robot vision cameras. It captures the data in near-time and processes it on ingest, making it quickly available within the KPA application for analytics and insight,” he said. “Our platform seamlessly processes, tags, geospatially locates, and indexes visual sensor data via the most sophisticated AI and machine learning available in the enterprise cloud ecosystem.”

Kespry collaborates with Microsoft

Kespry said it collaborated with Microsoft Corp. to ensure that Kespry Perception Analytics meets the rigorous needs of complex industrial operations. The system vertically integrates as an ISV solution for the Microsoft Dynamics 365 and Power Platform.

Kespry Perception Analytics runs on the Microsoft Azure cloud computing platform, which provides customers with the ability to integrate, manage, and deploy Kespry Perception Analytics at scale across their global networks using established tools and frameworks, said the companies.

“Kespry Perception Analytics is an excellent example of a business solution that combines next-generation image and video analysis via the power of Microsoft Azure cloud and Azure AI.” said James Phillips, corporate vice president at Microsoft. “Kespry’s solution requires incredible speed and network availability to ensure it can deliver on its compute demands and have the resilience and availability necessary as a business-critical application.”

Uses for KPA

Chemical, energy, pulp and paper, food and beverage, and port facilities capture petabytes of inspection data from handheld and fixed cameras, drones and other robots, and thermal guns. This data can provide critical insights into asset conditions and efficiency. But reliability and maintenance teams waste valuable time organizing and searching inaccessible, unaggregated, and difficult-to-parse data across multiple users and systems, said Kespry. This makes it difficult to proactively detect problems that can lead to asset failures such as leaks and corrosion.

Mathew provided an example of how Kespry Perception Analytics can benefit inspection applications.

“Energy companies require constant inspections of key assets such as pipelines, tanks, and stacks to maximize health and performance for 24×7 operations,” he said. “Kespry Perception Analytics solves one of the most significant, yet demanding problems energy companies face today: proactively and efficiently analyzing key insights from terabytes of visual data captured daily, and correlating the requisite operational data to identify and resolve issues.

“Similar to large industrial companies, ports have a host of assets that they need to regularly maintain, including dock components; cranes; tank farms; and infrastructure such as roads, bridges, and ladders. In order to inspect these assets, members of their engineering, maintenance and tenant services teams collect visual data on a daily basis. However, once that data is collected, there is no good way to store and organize it,” Mathew added. “Kespry Perception Analytics help ports become much more advanced in this process, enabling operations teams to better organize the visual data they collect, track trends over time, and tie visual data to specific assets in specific locations.”

Early users have provided support and feedback, he said.

“We are currently in the midst of major deployments with a pulp and paper giant, as well as one of the biggest port facilities in the country,” said Mathew. “Both organizations have been incredibly enthusiastic at the potential for Kespry Perception Analytics and see the long-term value for mitigating issues that could later become incredibly costly and difficult to address. The value of KPA is resonating with asset health and reliability teams, as well as data analysts within these companies.”