项目名称： 一种基于三维照明光场感知的分布式智能控光系统

A distributed light control system based on 3D lighting field perception

申报单位： 北京科技大学

综合介绍或申报理由：

1、需求分析 近年来，精细的近景照明在航天、航海以及智慧城市建设方面有越来越多的应用需求。

1）空间站应用需求 空间站地球飞行一周的时间约为一个半小时，因而表面环境光变化复杂多样。目前采用传统机器视觉系统进行航天器表面监测较为困难，环境光一旦发生变化，相关算法便会失效。

2）智慧城市建设需求 夜间条件下的视频监控是智慧城市建设中不可或缺的建设内容。夜间环境下，由于环境光线较弱、照明效果分布不均，导致难以进行图像处理算法的设计与应用。如夜间环境下人脸识别率较白天条件下识别率会低很多。

2、系统设计方法 提出一种分布式照明装置设计方法，通过三维成像环境感知技术，建立最优化目标函数，实现对单个光源的最优控制，最终达到作业面随动补光的目的，保证作业面光效的相对稳定。

3、解决的实际问题 1）看的清楚的问题 通过系统应用，照明光源出光效果随环境光变换而随动补光，进而保证使用人员不会产生眩晕、间歇性视功能障碍等问题。

2）算的稳定的问题 通过系统应用，保证了被拍摄区域光效稳定，确保了相关图像处理等算法计算的稳定性。

4、应用前景分析 本系统具有较强通用性和较好应用前景，如空间站、无人掘进机、水下机器人、智能交通、智慧城市等。

1. Requirement analyses Recently, the elaborated near-scene lighting has achieved more application requirements.

1) The space station application The space station is a low orbit space craft whose flight cycle around the earth is about one hour and a half. The light and shadow changes in its surface are complex. Currently, it is very difficult to employ the traditional computer vision methods to implement the surface abnormal state surveillance task because once the imaging environment changes the algorithm will stop work.

2) The intelligent city application The video surveillance at night is an indispensable part of intelligent city construction. In the night, since the environment light is weak and uninform, it will be very hard to use the image processing algorithms correctly; e.g., the face recognition ratio is low at night.

2. The system designing method A distributed lighting device is proposed. First, the 3D environment perception technique is developed. Second, an optimization function is built. Third, after the optimal computation the intelligent control of sole light source can be achieved. Finally, the dynamic light compensation can make the lighting effect in the operation surface stable.

3. The solution of practical problem Two problems can be solved by our system: to see clearly by people and to compute stably by machine.

4. Application prospect analysis The space station, the personless mining machine, and the underwater robot, etc.

项目宣传视频： https://b2b.alighting.com/headline/JQYVTSAB2P/videodetail-5477.html

主要技术参数：

本智能照明装置的实现可根据用户需求进行定制和调整，下面以所研发的原理样机为例，进行相关技术指标的介绍。

1）灯箱规格尺寸：1.0m*1.0m*0.3m；

2）LED灯珠规格：亮度350 mcd ~400mcd、色温6500K-7000K、白光、尺寸1.6mm*0.8mm、发光角120度~160度、响应时间100ms；

3）双目相机模块规格：1/4英寸传感器，3.0微米*3.0微米像元尺寸，380nm~1060nm感光谱段、USB3.0数据接口、YUV2560*720图像分辨率、30fps帧率、6.0cm双目相机基线；

4）LED灯控制电路模块规格：输入电压范围为8~450V、最大驱动电流1.5A，支持大于100颗LED灯珠同时工作；

5）工作环境：推荐工作环境温度范围-10摄氏度~30摄氏度、湿度范围40%~70%。

Main technique parameters The proposed system can be custom made for user. Currently a prototype device is developed, its technique parameters are listed below.

1) Size of light box: 1.0m*1.0m*0.3m

2) LED bead: luminance is 350 mcd ~400mcd; color temperature is 6500K-7000K; white color; size is 1.6mm*0.8mm; luminance angle is 120~160 degree; and response time is 100ms.

3) Binocular camera: size of sensor is 1/4; pixel size is 3.0um*3.0um; sensitometric range is 380nm~1060nm; data interface is USB3.0; image resolution is YUV2560*720, the frame rate is 30fps; and the distance between two camera is 6.0cm.

4) The control circuit box of LED: input voltage range is 8V~450V; the maximum drive current is 1.5A; this box supports more than 100 LED beads to work together.

5) Working environment: the recommend working temperature is -10~30 Celsius degree ; and the huity is 40%~70%.

与国内外同类产品或同类技术的比较情况：

目前暂未了解到国外有同类产品。 Not found currently.

经济评价分析：

本系统所设计分布式照明装置可根据用户具体应用需求进行元器件的选配。最低配置版本包括单一的照明单元和双目视觉相机；高配版本可包括任意多照明单元、双目视觉相机、其他环境感知传感器等。通过本系统的设计与应用，使得复杂光环境条件下的照明光效得到保证，可显著降低人员投入、提高拍摄质量、减少对昂贵微光传感器的应用需求等，具有较好的社会和经济效益。

This product can be constructed in different forms according to the practical application requirement of user. The mini-configuration can only include one lighting unit and one binocular camera; the complete-configuration can have any lighting units, the binocular camera, and other environment perception sensors. This system can provide the proper lighting effect, decrease the amount of labour force, improve the photography quality, and reduce the application requires of other expensive optics sensors.

技术及工艺创新要点：

主要创新点： 1）提出一种基于成像传感的环境感知分布式硬件系统设计方法与计算建模方法，能够实现对环境光场变化快速有效的定量化描述。

2）提出一种分布式照明单元的智能控制方法，设计了带约束的最优化函数，根据实际应用需求实现了单光源光效控制及整体照明效果优化调节。

The novelties of proposed system 1) A new designing and modeling methods of distributed lighting system which uses the imaging sensor is proposed. This system can realize the quantitative description of lighting field.

2) A new intelligent control method of lighting device is developed. The optimization function with the boundary conditions is built. And a kind of optimal lighting control can be achieved by considering both the lighting effect of single LED bead and the whole LED lamp.

实际运用案例和用户评价意见：

本系统核心技术的研发受多项课题支持，包括国家自然科学基金“空间站舱外作业机器人成像照明光场智能调控系统设计及建模方法研究”、北京辰达信科技发展有限公司“智能LED灯设计方法研究”、301医院“医院室内照明效果评价方法研究”等。

The development of this product has gotten supports by the National Natural Science Foundation of China and other enterprises’funding.

获奖、专利情况：

1）获奖 唯一发明人，第二十四届全国发明展览会“发明创业奖 项目奖”银奖，中国发明协会，2020。

2）专利 本产品目前已获多项授权发明专利，代表性专利如下： 唯一发明人，一种在轨航天器表面异常状况智能检测方法及其系统，中国发明专利，ZL201810174333.4，授权日：20191108. 唯一发明人，一种复杂环境光应用条件下的相机系统及像质调节方法，中国发明专利，ZL201811419276.8，

授权日：20191217. This product has been granted by several China invention patents and a Japan invention patent; it also got a sliver award by the China Association of Inventions in 2020.

是否有“智能设计”？

本产品综合采用了图像处理、计算机视觉、模式识别、物联网等相关前沿技术，实现了二维、三维环境光场的分析与感知计算，对出光单元的光效进行了精细控制，有效提高了作业面的照明效果，属于一款新型人工智能产品。

The proposed product utilizes the state-of-the-art techniques about image processing, computer vision, pattern recognition, and internet of things. Both the analysis and the perception of 2D and 3D light filed can be achieved, then an elaborated control of light unit can be performed. The proposed product can improve the lighting effect in the operation surface remarkable and belong to a kind of artificial intelligent product.

申报单位介绍：

北京科技大学，为原“211工程”建设高校、首批“985工程”优势学科创新平台建设高校；2017年，入选国家“双一流”建设高校；2018年，获批国防科工局、教育部共建高校。目前，学校已发展成为一所以工为主，工、理、管、文、经、法等多学科协调发展的教育部直属全国重点大学。 北京科技大学自动化学院成像环境感知课题组，长期从事相机标定、辐射诊断、图像处理、模式识别技术研究，近年来在光场调控、智能照明系统设计、智能光疗、智能激光控制等领域取得一系列成果。近五年，课题组在成像环境感知、智能照明系统设计、智能激光清洗、机器视觉系统应用等领域承担各类课题10余项，已发表论文近30篇，授权发明专利10余项，申报软件著作权10项。

产品图片：