CN-121980785-A - Intelligent scene illumination precise optimization method and system based on deep learning

CN121980785ACN 121980785 ACN121980785 ACN 121980785ACN-121980785-A

Abstract

The invention relates to the technical field of intelligent illumination, in particular to an intelligent scene illumination accurate optimization method and system based on deep learning, wherein a scene illumination database is established to cover basic parameters, scene characteristics, design schemes and user experience feedback and energy consumption data after operation of an illumination system; the method comprises the steps of evaluating illumination demand load based on user activities and environment light parameters, constructing an illumination optimization simulation space, carrying out dynamic operation simulation on a design scheme, loading the demand load, setting dynamic simulation acceleration time, simulating a full scene period in stages, and obtaining a precise optimization scheme through simulation results. The method and the system realize accurate assessment and dynamic adaptation of lighting demands through multi-dimensional data fusion and deep learning, expose design defects and optimize parameters in advance through virtual simulation acceleration verification, reduce actual deployment cost, support energy efficiency improvement of the whole life cycle of a lighting system and dynamic adjustment of user experience based on a data-driven self-adaptive optimization mechanism.

Inventors

ZHOU LIANG
JIANG JUNCHENG
ZHOU YUN
LUO JIEFU

Assignees

浙江百康光学股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260116

Claims (10)

1. An intelligent scene illumination precise optimization method based on deep learning is characterized by comprising the following steps: Establishing a scene lighting database, wherein the scene lighting database comprises basic parameter information, scene characteristic information, lighting design scheme information of a lighting system, user experience feedback information and energy consumption data after the lighting system operates; The scene characteristic information comprises user activity information and ambient light parameter information, and the illumination demand load is estimated according to the user activity information and the ambient light parameter information; Constructing an illumination optimization simulation space for the scene illumination based on the basic parameter information and the illumination design scheme information, and carrying out dynamic operation simulation on the illumination design scheme; Loading the lighting demand load in the lighting optimization simulation space, setting dynamic simulation acceleration time, performing operation simulation on a plurality of scene stages of the scene lighting based on the dynamic simulation acceleration time, and obtaining a precise optimization scheme through a lighting simulation result.
2. The intelligent scene lighting precision optimization method based on deep learning according to claim 1, wherein the estimating lighting demand load according to the user activity information and the ambient light parameter information comprises: dividing the scene demand stage according to the user activity information, and sequencing according to the demand intensity from high to low to generate a demand segment sequence; dividing the disturbance stage of the ambient light according to the ambient light parameter information, and sequencing the disturbance stages from high to low according to the disturbance amplitude to generate a disturbance segment sequence; The demand segment sequence and the disturbance segment sequence are arranged and combined, and the arrangement and combination results are ordered by integrating the intensity of demand and the amplitude of disturbance, so that an integrated demand sequence is generated; And setting a demand sensitivity threshold based on the scenerised lighting database, and carrying out lighting demand load assessment on the comprehensive demand sequence through the demand sensitivity threshold.
3. The intelligent scene lighting precision optimization method based on deep learning according to claim 2, wherein the lighting demand load evaluation of the integrated demand sequence through the demand sensitivity threshold comprises: Traversing the comprehensive demand sequence, and judging whether a demand item corresponding to the demand sensitivity threshold exists or not; if so, marking the demand sensitivity threshold value on the corresponding demand item, determining the ratio of the demand sensitivity threshold value to the demand sensitivity threshold value, and evaluating the lighting demand load according to the ratio; if the illumination requirement carried by the requirement item does not exist, selecting one requirement item, and judging the relation between the illumination requirement carried by the requirement item and the requirement sensitivity threshold; And if the integrated demand sequence is higher than the demand sensitivity threshold, marking the integrated demand sequence as an oversubscription sequence, and if the integrated demand sequence is lower than the demand sensitivity threshold, marking the integrated demand sequence as a low demand sequence.
4. The intelligent, scenerized lighting fine optimization method based on deep learning of claim 3, wherein said evaluating said lighting demand load according to said scale comprises: Obtaining a user expected comfort index, obtaining comfort performance of the lighting system on each demand item according to the demand sensitivity threshold and the proportion, and evaluating comfort stability according to the expected comfort index; setting a dynamic response threshold, wherein the dynamic response threshold is the maximum allowable delay time for switching the lighting system from the current lighting state to the target state, and switching the lighting system by adopting a requirement item higher than the requirement sensitivity threshold in unit time to acquire dynamic adaptability performance; The lighting demand load is evaluated based on the comfort stability and dynamic adaptation performance.
5. The intelligent scene lighting precision optimization method based on deep learning as claimed in claim 1, wherein the establishing a scene lighting database comprises: Collecting historical lighting system optimization data information, including lighting design schemes of various scene types; classifying the historical lighting system optimization data information according to the basic parameter information, the scene characteristic information, the lighting design scheme information, the user experience feedback information and the energy consumption data, and establishing a data index entry according to a classification result; And constructing a lighting design subset according to the lighting design schemes of each scene type, and establishing historical optimization effect tracking of each lighting design scheme for the lighting optimization simulation space to call data.
6. The intelligent scene lighting precision optimization method based on deep learning according to claim 5, wherein the data for the lighting optimization simulation space call comprises: determining the current scene characteristic information and user activity information, indexing the matched scene characteristic information and user activity information through the scene illumination database, and constructing simulation scene factors, wherein the simulation scene factors are digital representations of the scene characteristic information in the illumination optimization simulation space; determining basic parameter information and lighting design scheme information of a current lighting system, indexing the matched basic parameter information and lighting design scheme information through the scene lighting database, dividing the dynamic simulation acceleration time into stages, and setting different time acceleration factors according to scene requirement performance in each stage; And matching the scene lighting database according to the determined current scene characteristic information and the user activity information, and the basic parameter information and the lighting design scheme information of the current lighting system, tracking the corresponding user experience feedback information and the corresponding energy consumption data, and providing data support for the lighting simulation result.
7. The intelligent scene lighting precision optimization method based on deep learning of claim 1, wherein the running simulation of a plurality of scene phases of the scene lighting based on the dynamic simulation acceleration time comprises: Setting a time acceleration factor, which is used for defining the conversion ratio of the time length and the actual time in the actual operation of the simulation, so as to accelerate the simulation process to quickly obtain the optimization effect of long-term operation; defining a plurality of specific scene stages according to the actual scene characteristic information and the actual lighting design scheme information of the scene lighting, wherein each stage represents a scene requirement condition possibly served by a lighting system, and each scene stage corresponds to one time acceleration factor; loading corresponding user activity parameters and environment light parameters in each defined scene stage, and performing simulation operation; and (3) integrating simulation results of all scene stages, and evaluating the optimized performance of the lighting system in the assumed full-use period to obtain the lighting simulation results.
8. The intelligent scene lighting precise optimization method based on deep learning according to claim 1 or 7, wherein the precise optimization scheme is obtained through lighting simulation results, and the method comprises the following steps: deep learning is carried out on the scene lighting database, and the mapping relation between the basic parameter information, the scene characteristic information and the lighting design scheme information and the user experience feedback information and the energy consumption data is obtained after fitting; Setting a data updating period, carrying out data acquisition on the running performance of the deployed scenerification lighting system, recording acquired data into the scenerification lighting database, carrying out data updating on the scenerification lighting database according to the data updating period, and dynamically adjusting the mapping relation; And evaluating the optimization effect of the illumination simulation result according to the mapping relation to obtain the precise optimization scheme.
9. A deep learning based intelligent scenerising lighting precision optimisation system, the apparatus comprising a memory, a processor and a deep learning based intelligent scenerising lighting precision optimisation program stored on the memory and operable on the processor, the deep learning based intelligent scenerising lighting precision optimisation program being configured to implement the steps of the deep learning based intelligent scenerising lighting precision optimisation method as claimed in any one of claims 1 to 8.
10. A medium, wherein the medium has stored thereon an intelligent scenic lighting precision optimization program based on deep learning, and the intelligent scenic lighting precision optimization program based on deep learning, when executed by a processor, implements the steps of the intelligent scenic lighting precision optimization method based on deep learning as set forth in any one of claims 1 to 8.

Description

Intelligent scene illumination precise optimization method and system based on deep learning Technical Field The invention relates to the technical field of intelligent illumination, in particular to an intelligent scene illumination precise optimization method and system based on deep learning. Background With the development of intelligent architecture and green energy-saving technology, the application of a scene lighting system in the fields of office, home, public space and the like is increasingly wide. The traditional lighting system generally adopts a fixed illumination mode or simple sensor linkage control, is difficult to accurately adapt to dynamically-changed scene requirements, and causes significant problems of energy waste and unbalance of user experience. For example, in an office scene, the variation of natural light intensity along with time and the difference of human activity tracks can cause that a fixed lighting scheme cannot achieve both visual comfort and energy consumption efficiency, and in an intelligent home scene, the personalized requirements of different family members on illumination color temperature and brightness are difficult to meet through a traditional preset mode. In the prior art, although the partial illumination optimization scheme introduces the ambient light detection or behavior recognition technology, the partial illumination optimization scheme is dependent on single-dimension data, and lacks comprehensive modeling capability for scene characteristics and illumination system parameters. In addition, the traditional simulation method is generally based on static rules or empirical formulas, and cannot dynamically simulate the evolution process of the lighting effect under a complex scene, so that the optimization scheme is difficult to cope with sudden loads in actual operation, and is lack of deep learning capability on long-term operation data, so that the self-adaptive optimization of the whole life cycle of the lighting system is difficult to realize. The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present invention and is not intended to represent an admission that the foregoing is prior art. Disclosure of Invention The invention mainly aims to provide an intelligent scene illumination accurate optimization method and system based on deep learning, and aims to solve the technical problems that a traditional scene illumination system is difficult to accurately adapt to dynamic scene requirements, energy waste is caused, user experience is unbalanced, and full life cycle self-adaptive optimization cannot be realized due to the fact that a fixed illumination mode is adopted, single dimension data is relied on, static rule simulation is based, and deep learning capacity of long-term operation data is lacking. In order to achieve the above purpose, the invention provides an intelligent scene illumination precise optimization method based on deep learning, which comprises the following steps: Establishing a scene lighting database, wherein the scene lighting database comprises basic parameter information, scene characteristic information, lighting design scheme information of a lighting system, user experience feedback information and energy consumption data after the lighting system operates; The scene characteristic information comprises user activity information and ambient light parameter information, and the illumination demand load is estimated according to the user activity information and the ambient light parameter information; Constructing an illumination optimization simulation space for the scene illumination based on the basic parameter information and the illumination design scheme information, and carrying out dynamic operation simulation on the illumination design scheme; Loading the lighting demand load in the lighting optimization simulation space, setting dynamic simulation acceleration time, performing operation simulation on a plurality of scene stages of the scene lighting based on the dynamic simulation acceleration time, and obtaining a precise optimization scheme through a lighting simulation result. Optionally, the estimating the lighting demand load according to the user activity information and the ambient light parameter information includes: dividing the scene demand stage according to the user activity information, and sequencing according to the demand intensity from high to low to generate a demand segment sequence; dividing the disturbance stage of the ambient light according to the ambient light parameter information, and sequencing the disturbance stages from high to low according to the disturbance amplitude to generate a disturbance segment sequence; The demand segment sequence and the disturbance segment sequence are arranged and combined, and the arrangement and combination results are ordered by integrating the intensity of demand and the amplitude of distur