[Audio/Video] Hisilicon Hi3516CV610-MINI LINUX Development Board

The HI3516CV610-MINI 20S/10B Linux audio and video development board
is a new generation audio and video codec SOC from HiSilicon. It supports H.264, H.265, and SAVC 3.0 video encoding, with 4K@20/6M@30 encoding performance,
1T computing power NPU/motion detection intelligent analysis, and MIPI 2Line. Dual-channel sensor, low power consumption, fast start-up, supports AOV, Ethernet interface
 
 
.
Development board link: http://e.tb.cn/h.gHunxWXplKwqTB1?tk=VjUF3j3gCho
Compatible camera: http://e.tb.cn/h.gstFmdktVWfTQhm?tk=W6BV3j4wXtw
Development board SDK: https://gitlab.com/linux-kodo/hi3516cv610-kodo
Video tutorials (updated as needed): https://space.bilibili.com/390456922/channel/collectiondetail?sid=1280264
 
Key features:
⚫ 4K@20, 6M@30 resolution
⚫ Real-time binocular access, supports binocular models such as PTZ cameras
⚫ 1T computing power NPU, Transformer feature acceleration, large model edge deployment
⚫ SVAC3.0 Encoding standard, 20% improvement in compression rate
⚫ Smart Encoding 2.0, pixel upgrade, no increase in storage, 4MP 30-bit product can store 5GB per day
⚫ 32x exposure ratio and wide dynamic range, clearer backlighting in shops/lobby
⚫ Improved clarity, motion blur, face and license plate effects in low-light scenes
⚫ 15-meter voice pickup, 3-5 meter intercom, crying detection in home scenes
⚫ AOV low-power solution
⚫ OpenHarmony system and ecosystem
 
processor core
⚫ Supports ARM Cortex-A7 MP2
⚫ 950MHz clock speed
⚫ Supports 32KB I-Cache, 32KB D-Cache, 128KB L2 cache
⚫ Supports Neon and FPU intelligent engines
⚫ Neural Processor NPU
- Supports 1TOPs computing performance - Supports hundreds of mainstream algorithms in the industry, enabling customers to quickly commercialize - Supports Transformer feature acceleration, with a built-in dedicated multimodal large model - Tools support efficient model production and evolution - Built-in algorithms for face, human, vehicle, package, and pet detection, etc.
⚫ IVE2.5 upgraded operator, supporting motion detection, perimeter protection, tracking, perspective transformation, video diagnostics, and various intelligent analysis applications. AI ISP Processing
⚫ Supports basic ISP functions - Supports dynamic bad pixel correction - Supports 3A (AE/AWB/AF) functions, with user-adjustable parameters - Supports lens shading correction
⚫ Supports high dynamic range - Supports two-frame WDR fusion - Supports Advanced Local Tone Mapping - Supports strong light suppression and backlight compensation
⚫ Clarity and Noise Reduction - Supports fixed pattern noise removal (FPN) - Motion region enhancement algorithm - Supports temporal and spatial noise reduction - Supports multi-level motion detection
⚫ Color Management and Contrast Enhancement – ​​Supports Purple Fringing Removal 2.0 – Supports Key Color Enhancement – ​​Supports Local Dynamic Contrast Enhancement – ​​Supports Dehazing Functionality
⚫ PQ Tool 1.2