Leveraging Edge Video Analysis for Autonomous Machines to Improve Business Operations

The rapid adoption of autonomous machines or robots in today's industrial applications continues to rise. They are able to perform tedious, repetitive and complex tasks very smoothly with a high degree of autonomy. In a new World Robotics report from the International Federation of Robotics, over 2.4 million industrial robots are operating in factories worldwide. The evolution of new-age technologies has made integrated video analytics (IVA) possible that is now impacting a large number of applications across diverse environments, such as healthcare, transportation, buildings and factories.

While the capabilities of video analytics are relentlessly growing, businesses are increasingly gaining the ability to assess images and videos, detect and recognize objects and people, and derive actionable insights from what they see. In manufacturing, video analytics applications are assisting in increasing production, lessening energy consumption, and improving overall product quality.

The great enablers of video analytics which includes high definition video cameras provide information in real-time that can significantly enhance decision-making. However, in the case of autonomous mobile robots, AI-driven video analysis requires a vast amount of computational resources. Thus, the advancements in computing technology and AI algorithms have made it possible to perform edge video analysis (EVA). This is the ability to analyze the video on-location in real-time.

Edge video analysis leverages the power of microprocessor units (MPUs) combined with graphics processing units (GPUs) that exaggerate thousands of small processors to act AI-related tasks in a massively parallel way. The use of EVA is increasing rapidly as this system is making its way into a wide range of industries where real-time data is required promptly to track and detect obstructions, such as Oil and Gas, Transportation, Manufacturing, and more.

Numerous companies offer edge video analysis solutions to reduce human errors and minimize accidents. ADLINK Technology, for instance, is one such company that designs and manufactures a large number of products for embedded computing, test and measurement, and automation applications. The company's product line covers computer-on-modules, industrial motherboards, data acquisition modules and complete systems. It also offers IoT hardware and software, AI software and robotics solutions.

In the case of autonomous mobile robots, ADLINK's EVA solutions possess over enough computational power to control the robot, while evaluating its surroundings, feature ruggedized units that are ideal for use in a hostile ecosystem. A high-resolution camera augmented with ADLINK's AVA-5500 AI-enabled video analytics platform equipped with an EGX-MXM-P3000 GPU module, which is powered by an NVIDIA® Quadro® Embedded GPU, can detect issues on the railway track as far as a kilometer away.

So, while transferring everything to the Cloud that consumes a massive amount of bandwidth and also requires significant costs, edge video analysis performs local processing and shifting only germane information.