Over the past 40 years, machine vision has become an important tool for tasks such as defect detection on manufacturing lines, sorting and identifying various types of products, reading barcodes, and even improving safety compliance by monitoring whether workers are wearing appropriate protective gear. Much of this has been done using dedicated hardware and software designed for very specific scenarios.
Smartphones are disrupting this dedicated hardware model and greatly expanding the application of machine vision, while substantially reducing the cost and improving the performance of existing applications. Umajin is leveraging the continual improvement in smartphone hardware by providing efficient authoring of machine vision applications, as well as other Internet of Things (IoT) solutions that can be deployed on smartphones. The huge volume of smartphone production provides a cost and availability advantage that is hard to match with custom hardware.
Smartphones already have impressive specs with multiple high-resolution cameras, fast processors, GPUs, Wi-Fi and Bluetooth networking, and 5G/6G connectivity Combined with simple molded housings, flexible smartphone-controlled external LED lighting, and smart software, smartphones provide a very capable platform for machine vision applications.
For relatively simple tasks, such as identifying whether people on the factory floor are complying with safety gear requirements, an efficient Umajin application using the phone’s camera and processor/GPU can be sufficient to execute the task. More challenging tasks, such as detecting minute defects on products on a manufacturing line, may require the processing of many images at high speed. 5G/6G connectivity allows these tasks to be offloaded to cloud edge servers to facilitate very high throughput. Umajin provides unique machine vision tools and processing for these types of scenarios and orchestrates how these solutions scale from the device to the cloud, utilizing open standards and enterprise cloud services such as Microsoft Azure and AWS.
Smartphones also facilitate much greater flexibility than dedicated hardware. Umajin’s hardened runtime eliminates the need for binary updates, and allows applications running on the smartphones to be modified easily to incorporate improved AI models or to reconfigure a line for a new product. This ecosystem of low-cost ever-improving smartphones, high speed network connectivity, and enterprise cloud services, combined with Umajin’s Machine Vision building blocks and orchestration, makes for some exciting opportunities.