STMicroelectronics VD65G4 and VD55G4 are ultra-low-power 0.56-megapixel global shutter CMOS image sensors designed for battery-operated edge AI and always-on vision applications.
The main difference between the two sensors is that the VD65G4 features a color RGB Bayer pattern, while the VD55G4 is a monochrome sensor designed to capture visible to near-infrared (NIR) light. Both sensors use a compact 1/9-inch optical format and a 2.16 µm pixel pitch, utilizing Back Side Illuminated (BSI), CDTI, and 3D stacking technologies to achieve a tiny 2.73 x 2.16 mm bare-die footprint.
STMicro VD65G4 and VD55G4 specifications:
- Resolution – 0.56 MP (804 x 704)
- Chroma
- VD65G4 – RGB Bayer (RGGB).
- VD55G4 – Monochrome (Clear, Visible to NIR).
- Optical Characteristics – 1/9-inch (2.3mm) optical format with a 30° linear CRA and close to 1:1 aspect ratio.
- Pixel Technology – 2.16 µm x 2.16 µm pixel size utilizing global shutter, BSI, CDTI, and 3D stacking.
- Frame Rates – Up to 184 fps at full resolution, 271 fps at VGA, and 480 fps at QVGA.
- Dynamic Range – 68 dB.
- Output formats – RAW8, RAW10.
- Interfaces
- Data – MIPI CSI-2 (1-lane), I3C, or SPI (1-lane)
- Control – I2C (up to 1 MHz) or I3C (up to 12.5 MHz)
- Misc
- Hardware autoexposure
- Automatic dark calibration
- Noise reduction
- Defective pixel correction
- Gamma correction
- Background removal (image difference mode)
- 4×4 programmable frame statistics
- Context management (up to 4 contexts)
- Power
- VCORE – 1.1 V
- VDDIO – 1.2 V or 1.8 V
- VANA – 2.8 V.
- Consumption – 35 mW typical @ 60 fps, 1-2 mW in auto wake-up mode, 0.8 mW in standby
- Dimensions – 43-pin bare die in reconstructed wafer, 2.73 x 2.16 mm footprint.
- Temperature Range – -30°C to 85°C
The tiny, low-resolution sensors are optimized for wearables, AR/VR headsets, and smart home appliances. The sensors consume 35 mW typically at 60 fps, but more importantly, feature an “auto wake-up” mode that draws just 1 to 2 mW.
ST positions this as an ultra-low-power alternative to PIR sensors in smart IP cameras to reduce false alerts. In this sleepy mode, the sensor can autonomously detect motion, waking up the host MCU/SoC only when an event occurs. This allows the system to activate a primary, higher-resolution camera only when necessary, eliminating the need for inefficient 24/7 continuous video streaming.
To reduce the workload on the main processor, ST integrates several hardware features directly into the sensor via its 3D-stacked design, including autoexposure and noise reduction. The sensor can also handle tasks like generating 4×4 frame statistics, managing multiple settings (contexts), and removing background by comparing consecutive frames, all on-chip.
To simplify the development, the company is developing the CAM-65G4 and CAM-55G4 Promodule evaluation camera modules, as well as the VD65G4 and VD55G4 S-Boards, which it says will soon be available for online purchase.
The P-Board (STEVAL-CAM-M0I1) is a simple development kit that connects ST BrightSense camera modules. It features a MIPI CSI-2 output with I²C control and connects via an FFC cable for quick plug-and-play integration with systems like Raspberry Pi and STM32. The board supports interchangeable promodules with different sensors and lenses, making it easy to test multiple setups.
Additionally, the company offers the EVK Main (STEVAL-EVK-U0I1), a USB-based evaluation kit for testing the image sensors on a PC. It includes a main board with USB-C (USB 3.0) output and a mechanical holder. It also includes a quick plug-and-play setup with free STSW-IMG501 software for live streaming and testing. The kit supports both ready-to-use promodules and custom setups with M12 lenses via S-Boards, making it reusable across different development stages.
On the software side, ST is developing a full Software Development Kit (SDK) with drivers and dedicated evaluation software. Currently, the software is marked as “upcoming,” with only the hardware sensors and datasheets available. In the near future, developers can expect access to ready-to-use drivers for platforms like STM32 and Raspberry Pi, along with example code, open user manuals, and step-by-step tutorials designed to accelerate edge AI and embedded vision projects.
The VD55G4 and VD65G4 image sensors are currently listed in preview status and are not yet available through distributors. However, at the time of writing, the development boards are available for purchase, with the STEVAL-CAM-M0I1 P-Board priced at around $38.18 in small quantities, and the STEVAL-EVK-U0I1 EVK Main USB evaluation kit available for about $160.87. More details can be found on their respective product pages and in the press release.
Debashis Das is a technical content writer and embedded engineer with over five years of experience in the industry. With expertise in Embedded C, PCB Design, and SEO optimization, he effectively blends difficult technical topics with clear communication
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
