RV1109

Rockchip RV1109

Key features

• Dual-core ARM Cortex-A7 and RISC-V MCU
• 250ms Fast booting
• 1.2Tops NPU
• 5M ISP with 3F HDR
• Up to 3 sensors can be used for simultaneous input
• 5M H.264/H.265 encoder and decoder

The Rockchip RV1109 is a high-performance AI vision processor System on Chip (SoC) tailored for intelligent vision applications such as IP cameras. It incorporates a dual-core ARM Cortex-A7 32-bit processor with NEON and FPU capabilities. Each core has 32KB of instruction and data cache, sharing a 512KB L2 cache. The integrated Neural Processing Unit (NPU) supports hybrid INT8/INT16 operations, delivering up to 1.2 TOPs of processing power. Additionally, it ensures easy model deployment by supporting network frameworks like TensorFlow, MXNet, PyTorch, and Caffe.

Additionally, RV1109 features a new generation, fully hardware-accelerated 5-megapixel ISP and post-processing unit. It integrates multiple algorithm accelerators commonly used in IPC and CVR, including HDR, 3A functions (AE, AF, AWB), LSC, 3DNR, 2DNR, sharpening, dehazing, fisheye correction, gamma correction, and feature point detection, all operating in real-time. Equipped with two MIPI CSI (or LVDS/SubLVDS) interfaces and one DVP (BT.601/BT.656/BT.1120) interface, it allows for the construction of systems receiving video from three cameras simultaneously. The embedded video encoder supports 5M H.265/H.264 encoding and multi-stream encoding, enabling a 5M stream at 30FPS and a 720P stream at 30FPS to be processed concurrently. This setup allows high-resolution camera output to be encoded and stored locally, while a lower-resolution video is transferred to cloud storage in real-time.

DETAIL SPEC