Meraki Api Keys

[Harsh_Sharma]

Hi,

I wanted to participate in Cisco Devnet Competition (SMart India Hackathon 2020). 

I am not able to access the API Keys for the Meraki camera's. 

Any possible leads?

[ConnorL]

Hey Harsh, Connor here from Meraki Support.

What's the issue you're facing obtaining your API keys? As per our Docs, you'll need to enable API access by going to Organization > Settings > Dashboard API access, and then go to your profile (Click your email at the top right of any Dashboard page and click "My Profile") to generate a key.

See info here: https://documentation.meraki.com/zGeneral_Administration/Other_Topics/The_Cisco_Meraki_Dashboard_API

Kind regards,

--

Connor Loughlin
Network Support Engineer

.:|:.:|:. Cisco Meraki EMEAR 🇬🇧

For reference, many questions can be easily answered by searching our online documentation: http://documentation.meraki.com

[PhilipDAth]

This seems to be related to this program:

https://developer.cisco.com/site/sih/ 

It says:
"To design modules based on the above mentioned activities, video feed shall be provided to the student teams. These feeds shall come from the Meraki security camera. Typically, these shall come from the actual / imitated site i.e. schools, food serving area etc."

Do you have access to the Meraki org that is hosting the camera?  If so then follow the tip that @ConnorL  has supplied.

How do you plan to attack this problem?

Are you a student?

[Sreekar]

Hi @PhilipDAth @Harsh_Sharma 

I am working on the same problem.

@PhilipDAth  I don't have access to any org hosting the camera.

P.S: I am a student

[ConnorL]

You'll need a Meraki Dashboard account and a camera in order to utilise any API calls meant for MV. I'd reach out to the competition organisers to either obtain more information and/or a trial camera.

It's likely they'll need to give you access to a pre-configured Dashboard account. But this is separate from Meraki so you'll need to contact Cisco Devnet.

[Sreekar]

Hi @ConnorL 

On searching about how Meraki cameras work, I've found that they perform object detection on the camera and transfers the metadata over the APIs. Do they have any models built-in onto the on-board processor or they access the models through API? 

If the models are accessed through API, then if the internet connectivity is lost the cameras do nothing but recording the video.

[Sreekar]

And what if we want to load our own user-defined deep learning models (architectures), is that possible?

[PhilipDAth]

In the cameras themselves - that is not possible.

What you do is use the snapshot API for when a person is detected, download the image, and then run it through your own deep learning model.

[PhilipDAth]

>Do they have any models built-in onto the on-board processor or they access the models through API? 

Foir the kind of cameras mentioned, MV12, they have a people detection model built into the camera.  You have no access to the model.  You can only retrieve a snapshot picture (or video clip) when it triggers.

>If the models are accessed through API, then if the internet connectivity is lost the cameras do nothing but recording the video.

The cameras keep recording if the Internet is lost.  All processing is done on the camera.  Typically if you are using the people detection you use MQTT, which is sent locally from the camera to an MQTT server that you operate.