Modern manufacturing and logistics distribution centers need to increase the level of productivity by increasing the speed of delivery, expanding the scope of vision and work. However, this is constrained by conventional automatic detection and scanning cameras that use more traditional 4:3 or square machine vision CMOS image sensors, which have the serious disadvantages of redundant vertical pixels and lower frame rates.
The rise of new CMOS image sensors on the market, specifically designed for scanning applications, represents a major step forward for industrial applications. These sensors use state-of-the-art low-noise global shutter pixel technology to achieve optoelectronic performance that would have been possible with twice the size of a pixel just a few years ago. The small size and optical array matched to the packaged mechanical center can provide a powerful solution for the design of thin and light mobile cameras and modules, and enable compact mobile designs for small OEM barcode engines, mobile terminals, phone accessories, iot, contactless authentication systems, wearables, drones and robots.
Some of these are in a wide 2.4:1 format, which makes them particularly suitable for fast-moving objects or barcodes on wider conveyor belts. They can also seamlessly connect to CPU, ISP, DSP chipsets with a 2-channel MIPI CSI-2 interface at 1.2GHz.
CMOS image sensors for handheld/fixed and mobile barcode readers, OEM scanning engines and automatic recognition.
It’s not just a matter of size, scanning fast-moving objects requires a very short integration time to avoid blurring the captured image. Short integration times mean less photon accumulation, which then requires signal gain to produce the appropriate grayscale and contrast. There are additional optical aspects that further reduce the available photons (even before they reach the image sensor). Due to the wide range imaging required in some of these applications, a larger lens with F-number is required to ensure that the image is fully focused. A larger F-number means a smaller front stop, which reduces the available light reaching the sensor. As mentioned earlier, increasing the exposure time is not feasible, so if the sensor’s low optical signal-to-noise ratio is not good, then the only option is to increase the lighting, resulting in additional costs such as leds, power supplies, and stricter thermal management design to limit the increase in the sensor’s dark signal noise (typically doubling every 6 to 8 degrees of junction temperature increase).
This next-generation CMOS sensor addresses these low-light challenges with a small global shutter pixel feature and an “in-pixel” CDS (correlated double sampling), achieving a typical total readout noise of 3 electrons. In addition, the advanced dual-light guide design with a zero-gap microlens maximizes light coupling to the light-sensitive region of the pixel to achieve a good signal-to-noise ratio at low light. The light guide and microlens also enable a very wide angular response, closer to a backlit pixel (although these sensors are front-lit). These “in-pixel” optics can also reduce crosstalk between pixels, providing clear images.
Further discussion of image quality needs to include dark signals and their effect on fixed mode noise (FPN), especially at high ambient temperatures. High dark signal non-uniformity (DSNU) tends to result in more visible VFPN/HFPN in the image, even when using compensation techniques.
Application advantages of advanced industrial CMOS sensors
Currently, CMOS sensors for most bar code readers are around 1MP, and device manufacturers are competing on price, eroding profit margins. Thanks to their higher resolution and unique wide-scale format, the introduction of these new scanning-specific industrial CMOS sensors brings an upgrade that enables longer or wider scanning ranges. This is helpful in retail and logistics scanning where a larger area/field of view is required. End users will benefit from faster conveyor belts and greater throughput compared to existing solutions.
The compact size also makes these new sensors ideal for small OEM scanning engines and ultra-thin mobile platforms. The combination of full 2MP resolution, high frame rate and low power consumption ensures a compact scan engine with higher operating range and longer battery autonomy, ideal for handheld and mobile applications. Some sensors also embed specific barcode reading functions for fast scanning, such as fast wake up, enabling the decoding of the first frame within 10 milliseconds of power on.
Document scanning also requires a wide-scale format that is a particularly useful feature in mobile barcode terminals in the field of logistics and parcel delivery.
Future trends and applications
In addition to barcode reading, the fixed barcode scanners used in retail POS (point of Sale) are also gradually being used for object recognition, so that fruits, vegetables and hard-to-read labels can improve the customer’s shopping experience through self-checkout without interference. RGB sensors can provide enhanced object recognition, maintaining adequate spatial resolution for traditional barcode reading.
The new high-efficiency global shutter (typical value 1/PLS> 4000@660nm) sensor is ideal for general-purpose machine vision applications such as factory automation, automatic inspection, and more. In the past these applications integrated arrays in square format, but now the full HD format has become mainstream, and some of these sensors can do this by cropping or merging 2 x 2 pixels horizontally, allowing a single camera design to meet all the requirements.
CMOS sensors are also used in machine vision, robotics, and embedded imaging applications.
Robotic devices used to pick up and place goods or used to transport goods from one place to another use 2D barcode readers to identify goods or boxes and move through the factory floor. While 3D technology is often used for safety and collision avoidance, these new, smaller image sensors make it easy to design small stereoscopic 3D camera systems.
With the deployment of artificial intelligence and machine learning, the popularity of embedded vision systems has increased significantly. These sensors can also be the eyes of these systems.
Because of their small size, light weight, and low power consumption, advanced CMOS sensors can also be used in drones or remote iot cameras. The full HD image format is compatible with 16:9 displays, making these devices a natural fit for remote visual inspection or surveillance applications. Internal temperature sensing can even provide additional environmental information and/or calibration of temperature-related workpieces.
Due to their small size, sensor-based in-mouth 2D or 3D cameras will benefit dentistry. Similarly, color cameras for dental matching applications can also be implemented.
A related area that is growing is point-of-care scanning products. These are multi-functional handheld terminals that allow patients to perform medical examinations with their doctors via video call from home. The COVID-19 pandemic has significantly boosted this market due to isolation and home care needs. Cameras are an integral part of these terminals that need to perform a wide range of functions, such as zooming in on areas and monitoring the progression of scar tissue after surgery. The same camera also needs to complete other biometric and metrological tasks, and these scanning-specific sensors would be the ideal solution.