====== People Counting - Sighthound Video ======
===== Introduction =====
Due to required system upgrades, the prior Cognimatics (now Axis People Counter) solution was discontinued. A new Computer Vision based approach was used, based on the [[https://www.sighthound.com/products/sighthound-video|Sighthound Video ]] (SV) product. 

===== Improvements =====
Unlike the Cognimatics approach, SV uses computer vision to recognise objects within the camera view and classify them.

{{:classification_1.jpg?600|}}

We are also able to create a larger number of detection zones within the same camera frame, allowing a richer dataset to be generated. 
Counts will be broken into 5 minute segments allowing higher granularity of data.

===== Disadvantages =====
Because of the Computer Vision approach, a greater amount of computing power is required to process the vision data.