Towards window state detection using image processing in residential and office building facades

Abstract

Although building energy simulation is an essential tool for designers, numerous studies have proven the existence of a performance gap between the estimated and measured energy use. In response, researchers are exploring ways to collect authentic behavioural data to improve existing inaccurate data-driven occupancy behaviour models. An important occupant behaviour currently being studied is window operation behaviour as it results in large consequences on heating/cooling loads and ventilation rates of a building. Ex-situ camera-based window operation monitoring has been proposed as a solution. This study used image processing technologies that automatically identified windows on a façade and determined their individual state (i.e. open, partially open, or closed). The algorithm developed through this study yielded an 89% accuracy rate over all the windows tested. This algorithm was developed to specifically target punched façades with awning windows. Factors that affected the accuracy of ex-situ camera-based window operation monitoring included environmental conditions such as lighting, obstructions, and reflections. Furthermore, there were challenges in determining threshold values used to isolate important image data that defined the window state and location and identifying the significant peaks for window angle image data. The next steps for this research should determine appropriate threshold values that can be used universally through additional testing and to explore new image processing techniques for other window types.