A computer vision-based building space energy consumption control method and system

By combining computer vision and edge computing technologies with a lightweight crowd density estimation model, high-precision crowd status perception and dynamic energy consumption control within building spaces are achieved, solving the response lag and privacy concerns of traditional methods and improving the energy efficiency and user experience of public buildings.

CN122116273APending Publication Date: 2026-05-29CHINA YOUKE COMM TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA YOUKE COMM TECH
Filing Date
2026-02-12
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Traditional building energy consumption control methods suffer from slow response, limited coverage, and inability to accurately sense the status of people. Furthermore, multi-sensor solutions are complex, costly, and raise privacy concerns, making them difficult to promote in public buildings.

Method used

Employing a lightweight crowd density estimation model based on computer vision and an edge computing architecture, the system collects video stream data in real time through cameras, constructs a mapping relationship between zones and cameras, achieves high-precision crowd density perception, and dynamically generates device control commands for refined energy consumption control of zones.

Benefits of technology

It achieves non-contact, high-precision crowd status perception, simplifies system structure, reduces deployment costs, avoids privacy issues, provides low-latency real-time energy management, and improves energy-saving performance and user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122116273A_ABST
    Figure CN122116273A_ABST
Patent Text Reader

Abstract

The application provides a building space energy consumption control method and system based on computer vision, comprising the following steps: S1: partitioning a public area of a building space and constructing a correspondence relationship between the partition and a camera and an energy consumption device; S2: mapping the partition to a digital plane grid and constructing a mapping relationship between the grid and a camera field of view; S3: collecting and processing video stream data of each camera in real time, inputting the video stream data into a lightweight crowd density estimation model, and calculating a crowd density value in the field of view of each camera; S4: based on the real-time value of the crowd density value in the field of view of each camera and the mapping relationship between the grid and the camera field of view, a weighted fusion algorithm is used to calculate a crowd density grade of each partition; S5: according to the real-time crowd density grade of the partition, combining preset upper and lower threshold values and a control strategy, corresponding device control instructions are dynamically generated; and S6: the control instructions are sent to terminal devices such as lighting and air conditioning for execution, so that dynamic and fine energy consumption control of the partition is achieved.
Need to check novelty before this filing date? Find Prior Art