Virtual Light Sensing Technology for Fast Calculation of Daylight Autonomy Metrics.

Virtual sensing technology uses mathematical calculations instead of natural measurements when the latter are too difficult or expensive. Nowadays, application of virtual light sensing technology becomes almost mandatory for daylight analysis at the stage of architectural project development. Daylight Autonomy metrics should be calculated multiple times during the project. A properly designed building can reduce the necessity of artificial lighting, thus saving energy. There are two main daylight performance metrics: Spatial Daylight Autonomy (sDA) and Annual Sunlight Exposure (ASE). To obtain their values, we have to simulate global illumination for every hour of the year. A light simulation method should therefore be as efficient as possible for processing complex building models. In this paper we present a method for fast calculation of Daylight Autonomy metrics, allowing them to be calculated within a reasonable timescale. We compared our method with straightforward calculations and other existing solutions. This comparison demonstrates good agreement; this proves sufficient accuracy and higher efficiency of the method. Our method also contains an original algorithm for the automatic setting of the sensing area. The sDA metric is calculated considering blinds control, which should open or close them depending on overexposure to direct sunlight. Thus, we developed an optimization procedure to determine the blinds configuration at any time.

Image Synthesis Pipeline for CNN-Based Sensing Systems.

The rapid development of machine learning technologies in recent years has led to the emergence of CNN-based sensors or ML-enabled smart sensor systems, which are intensively used in medical analytics, unmanned driving of cars, Earth sensing, etc. In practice, the accuracy of CNN-based sensors is highly dependent on the quality of the training datasets. The preparation of such datasets faces two fundamental challenges: data quantity and data quality. In this paper, we propose an approach aimed to solve both of these problems and investigate its efficiency. Our solution improves training datasets and validates it in several different applications: object classification and detection, depth buffer reconstruction, panoptic segmentation. We present a pipeline for image dataset augmentation by synthesis with computer graphics and generative neural networks approaches. Our solution is well-controlled and allows us to generate datasets in a reproducible manner with the desired distribution of features which is essential to conduct specific experiments in computer vision. We developed a content creation pipeline targeted to create realistic image sequences with highly variable content. Our technique allows rendering of a single 3D object or 3D scene in a variety of ways, including changing of geometry, materials and lighting. By using synthetic data in training, we have improved the accuracy of CNN-based sensors compared to using only real-life data.