Nature of Disruption: INSPEKTO S70 Gen.2 includes 5000 built-in LEDs (light-emitting diodes) that are controlled using pulse-width modulation. They are arranged in different regions that enable the system to autonomously control the illumination direction and capture images with varying light direction and intensity. The images are then combined to construct a single reflection-less HDR (high dynamic range) image. The system can automatically control parameters like exposure, iris, illumination regime, and sequence of images used to create the final image. Additionally, it provides a vibration detection and resolution mechanism that detects objects running at a speed of up to 0.75m/sec. The INSPEKTO S70 Gen.2 is built on the principles of AMV-AI, a method of neural network architecture that mimics the human cognitive vision process. It includes three AI-based engines. The first is the acquisition AI engine to support vision tasks by adjusting the electro-optical imaging system’s operating parameters in real-time. The second is the detection and alignment AI engine to identify, classify, and perform 3D alignment of an object within the acquired image after a single sample input. It enables the system to recognize parts in a live video stream and determine the best moment to capture an image. The third AI engine in the system is responsible for inspection.

Outlook: Conventional machine vision solutions are built as a project from multiple components and can be applied to inspect only a specific manufacturing process. Inspekto claims that its AMV system contains both hardware and software and can be applied to inspect a wide variety of processes across industries. It device has a simple setup procedure that does not require training, data labeling, or annotation to create an immediate proof of concept (POC). Through the S70 Gen.2, Inspekto says it aims to offer a cost-effective quality assurance system for small and large manufacturing companies.