Device Strategies

Parameters	Description
Tracking Mode	Select the tracking mode of counting, including Heads Tracking and Feet Tracking: When the device detects both feet of the target in the FOV, it generates a trajectory line based on the movement path of the feet. When the target's head and shoulders are detected, a corresponding trajectory line is generated according to the movement path of the head and shoulders.
Detection Algorithm Setting	Select the detection algorithm according to the real applications. RGB+Depth: Suitable for most scenarios. RGB: Switch this mode when there are many false detections. Suitable for scenes with a large number of non-human objects mistakenly detected as people. For instance, the entrances and exits of a warehouse. Depth: Switch this mode when there are many false detections. Suitable for scenes with a large number of human-like objects. For example, a doll shop.
Record Track Start/Stop Points	Enable to record the start track points and end track points of people in the live view for the position adjustment of the detection line. It can store 5000 track points at most, with green as the starting point and red as the stop point.
Reset Cumulative Count on Schedule	Enable to periodically reset cumulative count on schedule. Cumulative Count includes: Total In/Out counting of each detection line. Max./Avg. Dwell Time of each detection region. Total Effective Audience and Avg. Attention Time of each attention region.

Tracking Mode

Select the tracking mode of counting, including Heads Tracking and Feet Tracking:

When the device detects both feet of the target in the FOV, it generates a trajectory line based on the movement path of the feet.

When the target's head and shoulders are detected, a corresponding trajectory line is generated according to the movement path of the head and shoulders.

Detection Algorithm Setting

Select the detection algorithm according to the real applications.

RGB+Depth: Suitable for most scenarios.

RGB: Switch this mode when there are many false detections. Suitable for scenes with a large number of non-human objects mistakenly detected as people. For instance, the entrances and exits of a warehouse.

Depth: Switch this mode when there are many false detections. Suitable for scenes with a large number of human-like objects. For example, a doll shop.

Record Track Start/Stop Points

Enable to record the start track points and end track points of people in the live view for the position adjustment of the detection line. It can store 5000 track points at most, with green as the starting point and red as the stop point.

Reset Cumulative Count on Schedule

Enable to periodically reset cumulative count on schedule.

Cumulative Count includes:

Total In/Out counting of each detection line.

Max./Avg. Dwell Time of each detection region.

Total Effective Audience and Avg. Attention Time of each attention region.