📢 Detection Features Survey

[Reolink Daisy]

Hey everyone!
We've received many requests for smart detection, and our dev team wants to collect your feedback on detection features for people, vehicles, and animals. Let us know your thoughts in the comments below:

1. How would you like to search for detected people events? By attributes like clothing, height, etc.?
2. Should we add options to search for vehicles by brand, color, etc.? Are categories like bicycles, motorcycles, buses, SUVs, sedans, trucks, and pickups enough?
3. Are there any specific animals you’d like us to detect?
4. What are your detection needs for daily home life (e.g., baby monitors, elderly care, bird feeders)?

We can’t wait to hear your ideas! The survey will close in 7 days.

[joseph_1979]

@reolink-daisy We need to have the event history for all detection objects for all cameras installed on our network. We just click on the event and the respective video should be played. Moreover, the customer may opt to play all the events between a specific period without the need to go through each camera. In addition, on the horizontal bar showing the time of the video playtime, we should have the objects detected and we can click on this and see the object rather than having to watch all the clip. We can have the number of objects detected in each clip and the time the object has been detected.

It would be nice to have the camera able to capture the number plate and be searchable (asking too much). Yes, it would be nice to search by categories but this entails for an intelligent AI detection. Way beyond what is available today. Today stationary cars are detected as moving, during the night persons are not detected, sometimes a swaying branch is detected as a person, etc. So one must ensure to optimise the existing AI algorithm.

Last year, if I recall correctly in July, we have been requested to reply to a survey on scene (now shortcuts). I have requested status and schedule but didn't hear anything. Has this survey been shelved?

[Matthew_635366143033465]

This post is deleted!

[user_654306224615431_654306224615431]

Regarding your survey I have two serious problems with your camera's
1/ Raindrops particularly at night are reported as vehicle if one is parked in view of camera
2/ Spider webs at night are reported as pet etc

On Duo 2 people or vehicle detection is not consistent across the whole field of view
All camera's do not detect people at long range (20Ft +), both Duo 2 & 12Mpix suffer this issue.

[Matthew_635366143033465]

I would like the option for smart detection (and clear) to trigger a webhook.

That webhook could carry a JSON payload with the various unique information -- whatever you choose to make available.

[CiprianB]

How would you like to search for detected people events? By attributes like clothing, height, etc.?
Attributes would be very helpful. Also, face detection for the household & perhaps friends. Imagine "family member detected" or "friend detected". And if the face is detected, "John detected".

Should we add options to search for vehicles by brand, color, etc.? Are categories like bicycles, motorcycles, buses, SUVs, sedans, trucks, and pickups enough?
Brand, color and definitely category. Also, license plate recognition would be nice. Imagine "foreign car detected" and "personal vehicle detected" (if wife is driving)

Are there any specific animals you’d like us to detect?
Most common pets. Imagine "pet detected" and "wild animal detected". maybe also specify the type of animal in the detection.

What are your detection needs for daily home life (e.g., baby monitors, elderly care, bird feeders)?
For me, it is a security system - so surveillance.

Kind regards,
Ciprian B.

[Jedi2020_32727772450952]

Color attribute is clearly very important for identifying objects which include clothing, hat, bike, car, and animals. Reolink should also seriously consider supporting more than 1 detection zone. For example, one detection zone for day and another detection zone for night. Push notification can then link to one or both zones. Many users have already requested this feature and some of the competitors already have this feature.