Aqara camera setup tutorial/explanation (for dummies). Part 1, "Detection Settings"

Aqara camera setup tutorial/explanation (for dummies). Part 2, "Function Settings" (Video Settings, Sound Settings, Privacy and Security) Hello, want to create this topic for users who bought an Aqara camera for the first time and often complain about the confusing setup of its options. As always, I give a lot of pictures to visualize the lesson.
The Aqara camera settings, by the development team, are made identically for all cameras (E1, G5, G350, G100). Once the user understands the setup algorithm, and in the future no one will have problems with any Aqara brand camera.
Like all devices, the settings window opens in this icon (…) in the upper right corner.
Fig.1.

The main settings are divided into two fields “Detection Settings” and “Function Settings”.
Fig.2.

The “Detection Settings” setting has three sections.
Fig.3.

The first section “Al Detection”, this is the main event detection setting. If there is a “shield” icon next to the event setting, this means that the event detection service works via a cloud service. I want to explain, the cloud service is used because the camera processor cannot process video information. This information is processed on the paid Aqara cloud service.
How to connect the cloud service, I described here Lesson 4. Setting up the APP security system (video surveillance cameras)
Fig4.

I will show pictures of the settings of the “Person detectipn” section. Similar settings for other functions. The settings contain sections for notifications, cloud recording, video analysis sensitivity, event detection region, re-search time. I show all the settings on the screenshots.
Fig5





The second section “Notifications”. This section duplicates the settings from the previous section, the peculiarity is that everything is collected in one place (for efficiency).
Fig6

The third section “Cloud Event Video”. Similarly, this section also duplicates the settings that are in the first section. These settings are also for the efficiency of setting up storage in the cloud service.
Fig. 7

In order not to overload with information, I will end this topic. In the next lesson I will talk about setting up the second block “Function Settings”.

11 Likes

That’s explained very well. Thank you. :clap:

But I still have a question about this: If I enable one of the detection features but disable notifications and cloud event videos for it, what exactly happens then?

2 Likes

Hello, I wrote, it’s just a duplication, a quick duplication.

2 Likes

Thanks for the quick reply. I understand that, but my question is:

What does this setting do (Enabled, but everything else is disabled)?

Compared to this setting (completely disabled)?

Is it the same thing, or is there a difference?

3 Likes

Theoretically, you will disable detection on the local storage scale (micro SD). I will give the answer for sure tomorrow.

6 Likes

Ah, you’re probably right about that. And only with active detection will anything show up in the timeline (log).


I’ve also been wondering if the following is true:

  1. Motion detection enabled with notifications enabled but pet detection disabled → as soon as a cat appears = notification

  2. Motion detection enabled with notification and pet detection enabled but notifications disabled → as soon as a cat appears = no notification

To me, that would make sense.
But I haven’t really been able to confirm it yet.
My cat is just camera-shy :rofl:


Edit: Okay, sorry. But I just figured it out myself.

It’s not what I thought. I just tried it with motion detection and person detection. The motion detection notification always comes through, even if you turn off the person detection notification. I thought that would distinguish between the two, but it detects both at the same time. So, if you have notifications enabled for both motion and person detection, you’ll receive two notifications at once. That’s too bad.

2 Likes

Purely from the logic of the algorithm, “motion” is the start for determining the event person, cat, box… If motion is not enabled, the camera algorithm will not start work on assigning the event name (person, face, cat…)
I am not developing this algorithm, this is my assumption of the recognition actions.

1 Like

Hi @Aqara_PM_Donie , can you tell us something about how this algorithm works? We would be very grateful.

1 Like

Each detection function in AI detection works independently, so which function you turn on will have corresponding results. If you only turn on the function and turn off cloud recording and message notifications, it can only be used to configure automation conditions and leave detection records in device logs.

The motion detection is usually frequent and of little value, so it defaults to a closed state. It is only guided by the app to be turned on when users paid Home guarding service, because the AI model in the cloud needs to analyze these short videos uploaded by motion detection and provide result description conclusions. Motion detection is not easy to miss events that occur, which is one of its major advantages.

6 Likes

Donie, thank you for the extended explanation of how the AI algorithm works.

1 Like