3TOPS Computing Power! Forlinx Embedded Launches the FET1126BJ-C: The New Generation AI Vision SoM Designed for Industrial Innovation
Forlinx Embedded and Rockchip jointly launch the FET1126BJ-C SoM, powered by the RV1126B processor. It features a board-to-board connector design, specifically developed for vision edge computing applications such as smart security and industrial inspection. It facilitates plug-and-play debugging during the R&D phase and enables quick maintenance of equipment in later stages.
The Forlinx Embedded FET1126BJ-C System on Module (SoM) features a 3 TOPS Neural Processing Unit (NPU) that supports mixed-precision inference with INT8 and INT4. This enables efficient execution of models like YOLOv5/v8 and ResNet for real-time tasks such as object detection and facial recognition on-device, enhancing response times and privacy without relying on the cloud.
The FET1126BJ-C features multi-camera input and a robust ISP image signal processor, providing clear and stable video output even in challenging lighting conditions, thus ensuring high-quality input for algorithms.
The FET1126BJ-C features AOA (Always On Audio) and AOV (Always On Video) standby modes, significantly lowering standby power consumption to the milliwatt range. This feature is specifically designed for applications with stringent battery life requirements, such as mobile inspection robots and battery-powered wireless monitoring devices.
The Forlinx Embedded FET1126BJ-C SoM features an LPDDR4 memory design, offering lower power consumption and a wider operating temperature range. Compared with the standard DDR4 temperature range of 0 °C to +70 °C, commercial-grade LPDDR4 chips can operate within an ambient temperature range of –20 °C to +85 °C, making them suitable for a wide variety of complex operating conditions.
The development board exposes all functional pins, offering a comprehensive suite of interface resources, including USB 3.0, MIPI DSI/CSI, HDMI, Gigabit Ethernet, UART, SPI, I2C, and a 40-pin GPIO header, facilitating flexible peripheral expansion and supports a wide range of development scenarios, such as embedded vision, edge computing, and AIoT applications.
The Forlinx Embedded OK1126BJ-C development board features a standard 40-pin GPIO connector, making it easy to connect to a wide range of external hardware, lowering the barrier to entry for development and accelerating the implementation of AI projects.
Forlinx Embedded offers comprehensive SDK support, covering Linux systems, drivers, middleware and AI toolchains, and comes pre-installed with mainstream frameworks such as TensorFlow Lite and PyTorch Mobile to accelerate model deployment. Whether you are an algorithm engineer or an embedded developer, you can quickly build prototypes and shorten time-to-market.
For industrial-grade applications requiring the deployment of vision AI computing power at the edge, the FET1126BJ-C SoM provides reliable underlying hardware support. Furthermore, to meet various physical design requirements, Forlinx offers the FET1126B-S SoM in a castellated-hole package, designed to meet development needs in compact spaces and high-vibration environments, thereby providing greater flexibility in hardware configuration for projects.
