摘要
室内光源亮度调节得好坏对室内设计及最终渲染效果会产生重要的影响,通常需要专业人士费时费力地手工调整.为此,提出了一种卧室光源亮度的自动调优算法.首先,在二维图像灰度熵的基础上进一步给出了三维光照熵的定义,以此作为评价三维卧室光源亮度优劣的数学模型;然后,利用模拟退火算法来求解卧室的最大光照熵值,从而得到光源的最优亮度值;最后,借助于Google AI采用的神经图谱(NIMA)图像美观度评估算法,验证了所提算法的正确性与有效性.该基于最大光照熵的光源的亮度自动调优算法也适用于客厅、餐厅等室内场景灯光亮度的设定与自动调优.
Adjusting the brightness of lights plays an important role in designing quality of interior layout and rendering effect of interior scene.Usually,professionals are required to adjust lighting brightness manually and skillfully.Thus,a novel method for optimizing bedroom lighting brightness is presented in this paper.First,the maximum lighting entropy(MLE)in 3D is defined for bedroom lighting brightness based on the image grayscale entropy(IGE)in 2D;this new mathematical model is built to evaluate the quality of bedroom lighting brightness.Then,the simulated annealing algorithm is employed to pursue the optimal lighting brightness of MLE for obtaining the best bedroom lighting effect.Finally,the neural image assessment algorithm recommended by Google AI is used to verify the correctness and effectiveness of our method through various bedroom cases.This method can also be extensively applied in other interior lighting such as living rooms and restaurants.
作者
豆雨桐
宋佩华
刘畅
贾金原
Dou Yutong;Song Peihua;Liu Chang;Jia Jinyuan(School of Software Engineering,Tongji University,Shanghai 201804;Department of Information Engineering,Nanning Normal University,Nanning 530001;School of Information Engineering,Nanchang Hangkong University,Nanchang 330063)
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2020年第8期1208-1217,共10页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金重点项目(U19A2063)。
关键词
光照熵
光源亮度
模拟退火算法
光照美观度
light entropy
light brightness
simulated annealing algorithm
aesthetics of light
