The surveillance devices store motion data and time series data on the cloud. There is infinite cloud storage available for each account. This data is kept secure for 90 days after which it is safely removed and destroyed.

The storage is secure and highly reliable.

A group of CloudCams and devices that are installed in an account form an intelligent grid. The ML algorithm that run on the surveillance data that is being captured in the cloud is able to track people and objects across cameras, gives the ability to stitch events across multiple cameras from historical data and more.

Surveillance Grid

The web UI provides a way to partition these CloudCams into several partition for ease of viewing.

Web UI

A group of CloudCam form an intelligent grid that is capable of capturing events that is covered by more than one CloudCam. For example, a person going out to meet someone is captured by a camera and that someone walking in the premises is captured by another camera. This event is stitched together in the Events tab.

Events Across Cameras

This feature can be very useful in quickly tracking down groups of people walking past several CloudCams.

The CloudCam platform comes with DHT22 which is a temperature and humidity sensor. Many such sensors can be easily integrated to the device if required.

Sensors time series data

Figure 1: A graph plotting temperature and humidity for last 12 hours using the data collected from a DHT22 sensor.

The whole paradigm of raising alerts and alarms changes with machine learning. The cameras learn over time about the movement, coverage, and timings of events, happening in front of the camera. The anomalies can be detected as abnormal events. These anomalies will be reported to you.

Here are the instant push notifications for any unusual events happening in front of your camera.

You get all the details of any abnormal events delivered to your mobile device.

The face detection and recognition are bundled as default features in CloudCam.

Face Detection

Face detection is the process by which the camera detects presence of faces in the view. In case it finds faces in the view it alerts the cloud. These faces will be shown in your alerts dashboard along with other alert information. To start with these faces will be nameless. The camera needs to be trained to recognize these faces. This process starts when you provide a name to a face. Initially it will be inaccurate as there is not much training data. As you add names to more faces the camera starts naming the faces that it finds further in the field.

CloudCam cloud can be configured to use available ML models. These can further work on the surveillance data and extract relevant information.
For example you can enable number plate recognition on your motion feed. As the name suggest Number Plate recognition allows the CloudCam to detect a vehicle’s number plate. This database of number plates is then available for searching a specific vehicle or tracking the movement of vehicles across CloudCams.

You can now share your motion feed or embed it in your website. This will make your business website more dynamic and will let your business better connect with your customers.

Following is the share from Lanco Lake View camera that takes hourly snapshots:

To create the share you need to login into the console. Then go to device settings page for whatever device you want to create a share. The form to create the share looks as below:

Motion Feed

CloudCam learns from the surveillance data that is being streamed to the cloud. This data is used to create the surveillance model of the CloudCam. This model helps in predicting abnormal events.