Annke C500 and its hardware and firmware. Any custom firmwares available for FH8656?

[alexk42]

I recently bought Annke C500 camera for testing. My initial goal was to make it to record to ftp server events.
I configured ftp server in Docker container. Tested it -works. Set in camera options and did a test and camera saved into the ftp folder a file named "test".
But it doesn't record the events to ftp server.
I was really unhappy with the camera and was thinking what a piece of junk I bought.
But then I have found SSH could be enabled. Then I ssh into the camera and find out it has busybox with commands.
I find out that it runs FH8656

https://www.fullhan.com/uploads/2021/11/163756337296075.pdf

Nice things it seems it has pretty decent specs. The fact that you can ssh into it makes it really nice. I wonder if there're nice community developed firmwares for that hardware.
The specs looks pretty good 64MB flash, 128Mb RAM it is better than my TPLink router running openwrt (flash 8mb, and 128mb ram).

Even though I like potential ability to make manythings with it. I really do not like what they have done. If the SoC supports up to 4Mp, how come the camera has 5mp. It seems that they just upscale to 5MP or even if they have a sensot 5Mp if the SoC doesn't support then what's the point of upscaling?

 

[alexk42]

To all owners of Annke cameras would you be able to post the details of you camera?
What is required to do, open camera portal using browser. enable ssh, reboot camera.
Ssh into camera ssh admin@cameraip
then execute command dmesg and please post the output here. I wonder if there're better cameras with better SoC than FH8656.

 

[alexk42]

Very interesting thing I found IP Camera SoC-Shanghai Fullhan Microelectonics Co.,Ltd. ( FHM )
FH8656 has face recognition optional

AI Engine 0.5TOPS computation power Human/Car detection Face detection/recognition (optional)

Would be cool if camera can identify that this is me and won't raise alarm.



The term "0.5T AI Power" in the context of an IP camera's System on Chip (SoC) specifications refers to the artificial intelligence (AI) processing capabilities of the hardware. Specifically, it means that the SoC can handle 0.5 tera operations per second (TOPS) dedicated to AI tasks.
Here's a bit more detail about what that can do:

1. Object Recognition: This level of processing power allows the camera to identify and track objects within its field of vision. This could include people, vehicles, animals, etc.
2. Motion Detection and Analysis: Advanced motion detection algorithms can be run to analyse movement within the scene. This is useful for security applications where anomalous movement might trigger an alert.
3. Image Enhancement: AI algorithms can be used for real-time enhancement of the image, such as noise reduction, clarity improvement, and color adjustment.
4. Face Recognition: In some instances, the SoC could handle facial recognition tasks, identifying specific individuals or detecting human faces in general.
5. Integration with Other Systems: The AI processing can be used to seamlessly integrate the camera with other smart systems, such as home automation, by responding to specific visual cues or patterns.
6. Efficiency: Having dedicated AI processing means these tasks can be performed on the device itself, reducing the need to send data to a central server for processing. This can lead to faster response times and lower bandwidth usage.

The figure "0.5T" gives you an idea of the processing capacity, and while it's a significant amount of computational power, it's typically suited for the aforementioned tasks rather than more complex or in-depth AI applications. Different IP cameras might have different AI power levels, and this would influence their capabilities and potential applications.