AcuSense של Hikvision

טכנולוגיית AcuSense של Hikvision מטביעה אלגוריתמים של למידה עמוקה במצלמות אבטחה נבחרות ומקליטי וידאו כדי לזהות באופן אוטומטי בני אדם וכלי רכב באמצעות ניתוח וידאו בזמן אמת ולדחוף התראות מיידיות למכשירים הניידים של המשתמשים.

על ידי מזעור אזעקות שווא המופעלות על ידי בעלי חיים, עלים נופלים, גשם כבד וחפצים נעים אחרים, AcuSense מאפשרת לעסקים ולבעלי בתים למקד את משאבי האבטחה באיומים אמיתיים.

בינה מלאכותית קלה ובמחיר סביר הגיעה לשוק ההמוני ולמגוון רחב של משתמשים. המשיכו לקרוא כדי ללמוד עוד.

Hikvision AcuSense technology features Classified Object Detection and Self-Learning Analytics.

Classified Object Detection
Self-Learning Analytics

In Hikvision’s laboratories, the deep-learning algorithms are trained to classify objects in videos into three categories – Humans, Vehicles, and Other. Users can then configure the algorithms to automatically detect humans, vehicles, or both based on their specific requirements, and then to notify them in real time.

Hikvision’s AcuSense technology pushes the use of Classified Object Detection to an ever wider variety of scenarios for users.

Hikvision AcuSense technology carries out four primary functions that can be used independently or combined based on the demands of the area to be protected.

Perimeter-Protection Analytics

Trespassers are detected at perimeters. Alerts get sent to users immediately via app notifications or short messages. Users can flexibly set the detection area and rules for customized protection.

Live Guard

Live Guard responds to trespassing events with a flashing light and customizable siren built into the security camera, alerting any would-be trespassers that they are being monitored and have already been spotted.

Quick Target Search

Video clips get sorted into “Person Files” and “Vehicle Files.” This allows homeowners and businesses to click a category and use time or location information to quickly locate the clip that they want.

There is no one-size-fits-all solution for video analytics. This is because in a real-world deployment environment there are often objects not covered by the training samples.

However, Hikvision provides Self-Learning Analytics with the Learn-by-Example approach to help users continuously improve the accuracy of the algorithms in specific application scenarios. Users mark the frequent false alarms generated within a particular scene, and the algorithms will learn from the examples and minimize similar errors in the follow-up rounds of alarm triggering.