What is Auto Tracking?
The world of smart security systems is advancing at a blistering pace. There are features we have today which were unthinkable just a few years ago. One such feature is the auto-tracking functionality for PTZ cameras.
Auto tracking implements complex AI algorithms that intelligently identify and track individuals and objects. These advanced algorithms closely integrate with the controls of the camera which allow it to pan tilt and zoom in concert to focus in on the target and get a consistent high-quality image and track it across the field of view.
This technology seems like it comes straight out of a spy movie. In reality, there are several distinct aspects that work together to provide the experience users are coming to expect. The first algorithm that needs to kick in is the object detection and recognition algorithm, which identifies the object that is intended to be tracked. After an object is detected further algorithms will predict the object’s movement direction and speed in order to preemptively move or zoom the camera accordingly. Without these algorithms the video would quickly lag behind the object, potentially losing it altogether. These algorithms are advancing rapidly, with frequent updates to their functionality, leading us to a new iteration of Auto Tracking.
What’s new in Auto Tracking 3.0?
Auto Tracking 3.0 is our latest version of auto tracking and is currently available on our new PTZ camera, the MTZ8250-IRAISMD-X. Auto Tracking 3.0 features some key technologies and features:
New Target Types
By training the AI systems on increasing numbers and types of target images, Auto Tracking 3.0 is able to track Aircraft and Ships, in addition to Humans and Vehicles, through the use of custom firmware.
PFA 3.0 focusing technology
This is the latest development of our PFA, or Predictive Focus Algorithm. This latest iteration further integrates intelligence and focusing algorithms to improve focus output during tracking.
Position Adaptive Algorithm
By using deep learning and statistical analysis, the new iteration of PAA is able to more accurately predict an object’s movement vector. This ensures the target remains in the center of the image and helps eliminate slowness or inaccuracies during tracking.
Interference Filtering Algorithm
Working in conjunction with PAA mentioned above, this algorithm similarly uses deep learning to predict direction and speed to improve target tracking in complex environments with occluding objects and dramatic lighting. This new IFA will better handle a tree being in front of a moving object, for example.
Speed Adaptive Algorithm
The Speed Adaptive Algorithm adjusts the rotation and zoom speed, as well as the exposure of the image to ensure a smooth track without jarring changes to the image.
Long Tracking Distance
By utilizing DPT technology to calculate the target frame more accurately, Auto Tracking 3.0 is able to track smaller targets at farther ranges than before. In ideal conditions, this can reach 1,500 meters.
Conclusion
Overall, Auto Tracking 3.0 is a significant improvement over its predecessors, providing users with an unparalleled level of tracking accuracy and versatility.
