Here is your first look at the editorial content for the upcoming Winter issue of Vision Spectra.
Embedded Vision
Embedded devices are in the midst of revolutionizing our industries, which some experts are beginning to call “Industry 4.0”. This era is where we solve industrial computing needs with networks of devices dedicated to specific or narrow tasks. The raw computing power we are capable of producing and powering inexpensively is creating new opportunities for the use of Artificial Intelligence on these “edge” devices.
Machine vision's interest in artificial intelligence mainly revolves around the use of a technique called deep learning which utilizes convolutional neural networks (CNNs) to classify images into categories, detect objects in images, or find unexpected anomalies in images such as damage on a produced part.
Why should you run AI on the Edge? While running AI on edge devices is challenging, decentralization of your image collection and processing has many benefits. With new technologies emerging and becoming ready-to-use products, Industry 4.0 is truly emerging.
Key Technologies: convolutional neural networks, deep learning, AI, edge machine vision
Machine Vision Lighting
Machine vision lighting technology must constantly evolve to meet increasing demands and advancing imaging technologies. Dynamic machine vision systems today, for example, leverage multi-sensor cameras and deep learning technology and therefore require intelligent lighting solutions. When designed properly, these systems solve previously unattainable factory applications while moving machine vision beyond the plant floor.
This article from Steve Kinney, Director of Engineering at Smart Vision Lights, details the latest innovations in lighting, including how optimized lighting can simplify deep learning vision solutions and the growing relevance of the Industrial Internet of Things, and how smart lighting fits into the equation. Additionally, the article will highlight real-life dynamic machine vision examples and the increasingly important role that lighting plays in system design.
Key Technologies: intelligent lighting, dynamic vision
Mobile Readers
From humble beginnings in the mid-1970s, the barcode and associated technology have come a long way, with a national day honoring barcodes declared in 2021. Today, barcodes are a key cog in the supply chain, with billions scanned every day. Mobile barcode readers are increasingly used, offering the industry a way to track items anywhere. For any barcode application, vision technology is vital in the form of light sources, sensors, controllers, and processing that together keep goods moving and track inventory. Innovations are needed, though, to meet mobile barcode reader demands.
Key Technologies: mobile barcode readers
Thermoelectric Cooling of IR Cameras
The use of outdoor cameras has seen a massive influx with the rise in security monitoring by government and private security agencies. Security cameras are installed with an objective to reduce crime or improve public safety. CCTV and IP security cameras are installed on every corner in cities, outside hotels and retail stores, all-around entertainment venues, stadiums, and other commercial and industrial properties. Thermal cameras (infrared sensing) are now often used to improve facility and border security at night. Most recently, thermal cameras have been implemented to detect people’s body temperatures as a security and safety measure in the fight against COVID-19. No matter the camera technology implemented, it is critical that the optimum operating temperature of sensitive imaging components is maintained during use to ensure high-quality images. Active thermoelectric coolers utilizing the Peltier effect offer advanced thermal management solutions. These thermoelectric devices meet the size constraints and high-temperature rating required for optoelectronics implemented in outdoor security cameras.
Digital cameras use two main types of imaging sensors: CCD (charge-coupled device) sensors and CMOS (complementary metal-oxide-semiconductor) sensors. Both sensors convert light (photons) into an electrical charge (electrons) using an intricate 2-D array of photodetectors (pixels). These individual buckets of charge are then amplified and digitized to create the digital image. The difference between the two main types of sensors is how and where this is accomplished.
Many cameras are now implementing thermal imaging technology for detection at night. These cameras create an image using infrared radiation (heat), similar to a common camera that forms an image using visible light. Instead of the 400- to 700-nm range of the visible light camera, infrared cameras are sensitive to wavelengths from about 1,000 nm (1 μm) to about 14,000 nm (14 μm). Thermal imaging cameras use focal plane arrays (FPAs) that respond to longer infrared wavelengths. Some smart thermal cameras include powerful video processors with access to the raw thermal video as it leaves the imager for more accurate detection. Advanced video processing also gives thermal cameras extended range and coverage, detecting human-sized targets at ranges exceeding 600 meters. Thermographic cameras can be broadly divided into two types: those with cooled infrared image detectors and those with uncooled detectors. Cooled detectors exist to maximize detection performance and viewing range.
Key Technologies: CCD, CMOS sensors, thermal imaging, IR, thermoelectric cooling