Manufacturers use QualityVisor to accelerate production and improve product quality by performing real-time visual inspection using computer vision (CV) and machine learning (ML). QualityVisor is a solution offering and deployment includes selection and provisioning of cameras, lights, edge software, and cloud integration, delivered on-site quickly and efficiently, making use of Amazon Lookout for Vision to create ML models capable of spotting anomalies from a live production line with as few as 30 training images.

Manufacturers competing in the global economy must leverage technology to improve product quality without impacting production speed. Traditional approaches to product quality inspection are unreliable, slow, expensive, and even dangerous because they rely on a human operator to manually spot check products as they move through the production line. Incorporating a custom computer vision solution is generally slow and expensive, which makes it difficult to scale out the solution across a portfolio of product lines at multiple locations.

QualityVisor by TensorIoT is machine learning enabled visual product quality monitoring solution that leverages pre-configured hardware and software that runs at the edge and integrates into the mature ecosystem of AWS cloud services. The hardware components, such as cameras, lights, and edge compute devices, are quickly and easily packaged together to assemble a custom system to satisfy the unique on-site environmental factors. The solution runs on-site, ensuring the system can address low latency and privacy requirements, while also saving money on cloud storage and network transmission. Through this consulting offer, TensorIoT will work with you to deploy QualityVisor, including the selection, installation, and tuning of the software and ML models for Amazon Lookout for Vision.

QualityVisor by TensorIoT for product quality monitoring is usually implemented in two phases:

Phase 1: Assessment and Prototyping In this phase, TensorIoT will work together with you to 1) define project scope, 2) assess the site environment for unique factors that may impact the implementation and performance and 3) build and demonstrate a working solution prototype. A working prototype that meets your requirements will demystify many concerns including solution performance metrics, production deployment timeline, TCO (total cost of ownership), and scalability. This first consulting phase takes about 4 weeks to complete with costs starting from US$30,000 plus reimbursement for any hardware, equipment, and incidentals. Deliverables include the physical prototype, prototype performance report, and the bill of materials for the prototype build.

Phase 2: On-Site Deployment Pilot Contingent upon a successful completion of Phase 1, TensorIoT will deliver the prototype to your site and redeploy it onto the production line. After installation, TensorIoT will monitor the solution performance and remediate any potential issues as they arise to ensure the solution continues to meet your requirements in production. This second consulting phase takes 8 weeks of elapsed time with costs starting from US$20,000 plus reimbursement for any hardware, equipment, and incidentals. Deliverables include a detailed report on TCO, hardware installation Instructions, software configuration, and AWS runbook.

Together, TensorIoT and AWS provide powerful holistic IoT, machine learning, artificial intelligence, and other advanced solutions for manufacturers and industrial companies. With AWS as the foundation and guidance from top TensorIoT experts, machine learning tools can deliver a wide range of benefits to your company and improve your business workflow.