ABSTRACT
Natural ventilation plays an important role in human comfort in indoor environments. In order to accurately estimate natural ventilation rates, the window opening area must be known. While sensors do exist for detecting window state, they are neither universal nor informative of the extent to which windows are open. This motivates the use of photographic and machine learning techniques to predict window state; however, without a high quality dataset, training such a machine learning model would be unlikely. This work details the creation of a hybrid dataset composed of real and rendered images for window state prediction. We use Rhino/Grasshopper daylight simulation and ray-tracing to generate photo-realistic renders to combine with a manually-captured photographic dataset.
- Ertugrul Bayraktar, Cihat Bora Yigit, and Pinar Boyraz. 2019. A hybrid image dataset toward bridging the gap between real and simulation environments for robotics. Machine Vision and Applications 30, 1 (01 Feb 2019), 23–40. https://doi.org/10.1007/s00138-018-0966-3Google ScholarDigital Library
- Tomas Hodan, Vibhav Vineet, Ran Gal, Emanuel Shalev, Jon Hanzelka, Treb Connell, Pedro Urbina, Sudipta N. Sinha, and Brian Guenter. 2019. Photorealistic Image Synthesis for Object Instance Detection. https://doi.org/10.48550/ARXIV.1902.03334Google ScholarCross Ref
- Anna Klimkowska, Stefano Cavazzi, Richard Leach, and Stephen Grebby. 2022. Detailed Three-Dimensional Building Façade Reconstruction: A Review on Applications, Data and Technologies. Remote Sensing 14, 11 (2022), 2579. https://www.proquest.com/scholarly-journals/detailed-three-dimensional-building-façade/docview/2674390438/se-2?accountid=14656Google ScholarCross Ref
- Yuzhen Peng, Yue Lei, Zeynep Duygu Tekler, Nogista Antanuri, Siu-Kit Lau, and Adrian Chong. 2022. Hybrid system controls of natural ventilation and HVAC in mixed-mode buildings: A comprehensive review. Energy and Buildings 276 (2022), 112509. https://doi.org/10.1016/j.enbuild.2022.112509Google ScholarCross Ref
- Iftikhar A Raja, J.Fergus Nicol, Kathryn J McCartney, and Michael A Humphreys. 2001. Thermal comfort: use of controls in naturally ventilated buildings. Energy and Buildings 33, 3 (2001), 235–244. https://doi.org/10.1016/S0378-7788(00)00087-6Google ScholarCross Ref
- O Seppänen and W J Fisk. 2002. Association of ventilation system type with SBS symptoms in office workers. Indoor Air 12, 2 (June 2002), 98–112.Google ScholarCross Ref
- Alexander Sorokin and David Forsyth. 2008. Utility data annotation with Amazon Mechanical Turk. In 2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops. 1–8. https://doi.org/10.1109/CVPRW.2008.4562953Google ScholarCross Ref
- J Sundell, H Levin, W W Nazaroff, W S Cain, W J Fisk, D T Grimsrud, F Gyntelberg, Y Li, A K Persily, A C Pickering, J M Samet, J D Spengler, S T Taylor, and C J Weschler. 2011. Ventilation rates and health: multidisciplinary review of the scientific literature. Indoor Air 21, 3 (Feb. 2011), 191–204.Google ScholarCross Ref
- Jeremy Wong, Julian Donges, Andrea Gasparella, and Adam Rysanek. 2023. A training dataset for machine learning-based prediction of window opening position in a naturally ventilated building, In The 11th International Conference on Indoor Air Quality, Ventilation & Energy Conservation in Buildings (IAQVEC2023), R. Ooka (Ed.). E3S Web of Conferences 396, 02034. https://doi.org/10.1051/e3sconf/202339602034Google ScholarCross Ref
- Yinda Zhang, Shuran Song, Ersin Yumer, Manolis Savva, Joon-Young Lee, Hailin Jin, and Thomas Funkhouser. 2016. Physically-Based Rendering for Indoor Scene Understanding Using Convolutional Neural Networks. https://doi.org/10.48550/ARXIV.1612.07429Google ScholarCross Ref
Index Terms
- On the Creation of a Hybrid Dataset to Predict Window States
Recommendations
Illumination from Shadows
In this paper, we introduce a method for recovering an illumination distribution of a scene from image brightness inside shadows cast by an object of known shape in the scene. In a natural illumination condition, a scene includes both direct and ...
Appearance Derivatives for Isonormal Clustering of Scenes
A new technique is proposed for scene analysis, called "appearance clustering.” The key result of this approach is that the scene points can be clustered according to their surface normals, even when the geometry, material, and lighting are all unknown. ...
Fast image synthesis of virtual objects in a real scene with natural shadings
We propose a new method for superimposing synthetic objects with natural shadings and cast shadows onto a real scene whose illumination condition is dynamically changing in this paper. In general, high computational cost for rendering virtual objects ...
Comments