I’m new to all of this having come from a traffic background so am trying to navigate through the different models Cloud Stream vs Cloud API.
Take the scenario:
There are several hundred IP cameras plugged into the city’s fibre / traffic network monitoring the streets - many are HD, but not all. We already bring images back to our VMS in the office but would like to run OpenALPR over the image set to help with license plate / make / model analysis.
Of the main 50 cameras covering the road network, lets assume there are around 10,000 passing vehicles per day, per camera… so about 15million vehicle movements per month. With the cloud stream basic at $39/camera/month this gives a monthly fee of $1,950/month. With CloudAPI it would be 15 x $1995… so almost $30k per month for what looks like essentially the same outcome… I can’t think that the hosting of a cloud service is that more expensive is it?? Anyone else confused? What would you choose?
The Cloud Stream service connects to live running video streams and will send a plate result for every license plate that it sees. It’s mainly geared towards surveillance applications, and it is priced based on number of cameras and not by volume of data.
The Cloud API service accepts still images and reports any plates found in each image. It is geared towards application integration and is priced by number of images rather than number of cameras.
If you’re doing still image recognition at high volumes, the recommended service is Numerus (https://rekorsystems.com/solutions/numerus/) which is about 1/3rd the cost that you calculated, however still not as cost-effective as Cloud Stream in this case.