When designing a solution that requires the use of an active pixel array CMOS digital image sensor, a large number of sensor specifications must be considered. For example, the sensors resolution, optical format, shutter type, maximum frame rate, dynamic range, signal-to-noise ratio (SNR), and pixel architecture, among others. Adding to the complexity are the sensors features/functions, such as power consumption, interface, package type, on-board HDR processing, and regions of interest. The optimal choice is not always immediately clear.

To help filter through these specifications and features, one important consideration is the sensors intended application. Some applications require very high resolution to capture stationary objects, while others need to detect fast-moving objects and be able to reproduce a "stop-action" effect. Another important application consideration is power consumption requirements. In fixed installations, the sensors power consumption may not be critical. However, in portable applications where the sensor must operate on battery power, the sensors energy efficiency becomes crucial.

When selecting a sensor, the most appropriate starting point is the speed of the application, i.e., how fast the object is moving. This will determine the required shutter type. In the realm of digital image sensors, there are primarily only two choices: rolling shutter and global shutter.

Image Sensor Considerations

Rolling Shutter

Digital image sensors are arrays of pixels arranged in rows. With a rolling shutter image sensor, each row in the array is exposed sequentially from the top to the bottom of the array. In other words, the exposure times of adjacent rows are slightly different (known as the row time), with a time difference of approximately 10 microseconds between adjacent rows.

In stark contrast to rolling shutter, a global shutter exposes every pixel in the array simultaneously. These sensors must have pixels with a "storage node" capable of storing charge during the readout of the entire sensor. Both rolling shutter and global shutter have their respective advantages and disadvantages.

Compared to global shutter, rolling shutter is more cost-effective and easier to implement. In global shutter sensors, the storage node is susceptible to stray light, so noise tends to be higher. Additionally, the storage node sits adjacent to the pixel, imposing constraints on pixel size. Correspondingly, the disadvantage of rolling shutter is its susceptibility to motion artifacts when capturing fast-moving objects.

Because rolling shutter arrays expose rows sequentially, spatial distortion occurs when photographing moving objects. Similarly, since different regions of the array are captured at different times (potentially under different lighting conditions), they can also be affected by unrelated illumination.

Thus, rolling shutter arrays perform poorly for capturing moving objects but are an excellent choice for static, high-resolution applications.

Global Shutter

Global shutter is suitable for scenarios where rolling shutter performs poorly, including fast-moving objects, especially those with high angular velocity. Applications where global shutter excels include augmented vision (AV), virtual reality (VR), machine vision (MV), and any environment with high vibration, such as barcode scanners and robotics applications.

Global shutter also offers other advantages: since the entire array is exposed simultaneously, the global shutter can be directly synchronized with other global shutters or light sources (such as flash). Global shutter also makes automatic exposure control easier to implement, as it does not need to deal with unrelated illumination.

Global Shutter Performance Considerations

The first step in evaluating global shutter performance is to refer to its demonstrated specifications. When evaluating global shutter sensors, both application-oriented metrics and performance-oriented metrics need to be considered. Application-oriented metrics help you select the right sensor within a specific product family for a given application, while performance-oriented metrics help you compare products from different manufacturers.

The simplest approach is to start with application-oriented metrics. The five main metrics determined by the end-use are as follows:

1. Resolution
2. Optical Format
3. Global Shutter Efficiency (GSE)
4. Frame Rate
5. Power Consumption

The priority of these five metrics will vary depending on the specific requirements of the end-use. For example, a high-resolution application might require a resolution of 2 megapixels (MP) and an optical format of 1/2.8 inch, while a low-resolution application might only need VGA resolution and a 1/8 inch optical format.

Frame Rate

Frame rate is measured in frames per second (fps) and indicates the number of images a sensor can capture in one second. Higher frame rates are required when capturing fast-moving objects to avoid blur.

Global Shutter Efficiency (GSE)

As mentioned above, GSE is a ratio that indicates the global shutters ability to suppress stray light. It is typically specified at a particular wavelength and f/stop setting. Higher values indicate better performance.

Energy Efficiency

Low-power optimization in consumer applications is crucial during development, especially for applications such as augmented reality (AR), virtual reality (VR), and mixed reality (MR) headsets. Additionally, autonomous mobile robots (AMRs) and handheld barcode scanners are a few examples of battery-powered devices in the industrial space. By improving the energy efficiency of these devices, their operational life can be significantly extended, reducing charging frequency and improving overall user experience. The remaining metrics are performance-oriented and can be used to compare products from different manufacturers.

Signal-to-Noise Ratio (SNR)

Signal-to-noise ratio is measured in decibels (dB) and specified as a maximum value. It is a measure of the sensors performance in low-light (i.e., small signal) situations. Higher values indicate better performance. The maximum SNR (SNRmax) is a true reflection of the linear full well (LFW), or essentially the number of photons a pixel can capture.

Dynamic Range

Dynamic range, also expressed in dB, represents the ratio of the maximum measurable input signal to the minimum measurable input signal (i.e., the noise floor). It indicates the sensors ability to handle varying input signal intensities within the same scene. Higher dB values are better. A tunnel is a good example of a scene requiring high dynamic range, as the interior of the tunnel may be dark while the exterior light is bright. The sensor needs to be able to adapt to both conditions within the same scene.

Beyond performance metrics, certain applications may require the sensor to have specific features to perform particular functions or possess unique capabilities. Not all global shutter sensors come with these features. The application requirements will determine which features are needed and which sensors can be considered.

Synchronized Sensor

One of the advantages of global shutter exposing the entire sensor array at once is that the moment of array exposure can be precisely synchronized with other events, such as other sensors and flashlights.

Through the synchronized sensors "trigger" mode, the flash can be controlled for precise active illumination, or multiple cameras can be synchronized for stereo or wide-screen capture.

Embedded Auto Exposure

The auto exposure function enables the sensor to automatically control gain and exposure based on given lighting conditions. Auto exposure is a fundamental function for the sensor to adapt to dynamic lighting conditions.

By embedding this function directly onto the sensor, exposure control can be faster, allowing real-time response, whereas relying on host control results in slower response. Embedded auto exposure is essential for most high-speed applications.

Scene Switching

The scene switching function allows the sensor to quickly change settings based on different resolutions, gains, exposures, and frame rates to adapt to different imaging scenarios. In many sensors, these "scenes" are stored and can be changed dynamically with a single register setting.

Programmable and Switchable Regions of Interest (ROI)

Regions of interest in an image are relevant collections of pixels, primarily used for object resolution. ROI allows the sensor to focus on a specific area by filtering out other parts. This is a method to optimize data transfer and processing. Programmable ROI enables real-time computer vision applications.

In summary, application-oriented metrics, performance-oriented metrics, and specific features can be combined to help select a specific global shutter sensor from a range of sensors and determine the sensor manufacturer that meets the end-use requirements.

Hyperlux SG Series Global Shutter Sensors

onsemi has developed a family of high-performance, small-size global shutter sensors called Hyperlux SG, including the ARX383, AR0145, and AR0235. The Hyperlux SG series sensors combine industry-leading global shutter efficiency (GSE) with low-power operation, making them ideal for portable, high-vibration applications.

Figure 1 Hyperlux SG Image Sensor Family

The Hyperlux SG series employs a novel and innovative global shutter pixel design optimized for accurately and quickly capturing motion scenes. It delivers clear, low-noise images in both low-light and bright scenarios.

The Hyperlux SG sensor series features:

• Horizontal / vertical mirroring, windowing, and pixel binning
• Programmable regions of interest (ROI)
• On-chip trigger mode for synchronization
• On-chip auto exposure
• Built-in flash control
• Scene switching
• Flexible control of row skip and column skip modes

The sensor portfolio ranges in resolution from VGA (640 x 480) to 2.3 megapixels (1920 x 1200), optical formats from 1/8 inch to 1/2.8 inch, and frame rates up to 120 frames per second, suitable for various high-speed imaging applications. Each specification has advantages over comparable products, and it is the combination of superior performance and features that makes the Hyperlux series sensors unique in the market. These sensors are ideal for high-speed applications including barcode scanning, machine vision, AMR, AGV, AV/VR/MR, drones, and 3D scanning.

Figure 2. Hyperlux SG Application Areas

To facilitate product education and system design, a comprehensive development platform is also provided to enable rapid system development. It includes complete testing capabilities to evaluate the product prior to design and provides reference designs for use during the design phase.

Global shutter sensors are the optimal choice for high-speed image applications. Once the shutter type is selected, there are still various specifications and features to choose from to ensure the sensor is suitable for the intended application.

When evaluating global shutter sensors, one important consideration that must always be kept in mind is global shutter efficiency (GSE). Without a sufficiently high GSE, all other specifications combined may still produce unacceptable motion artifacts in the image. For applications demanding low power, high performance, and high GSE, onsemi has developed the Hyperlux SG series of global shutter sensors.