Evaluation of light measurements for indoor and outdoor classification using neural networks

The objective classification of outdoor time has the potential to benefit applications involving the effect of outdoor exposure on various health outcomes such as happiness, stress, or myopia. The focus of this work is the use of different combinations of multiple light measurements as inputs to an artificial neural network (ANN) to classify indoor and outdoor environments. Seven different light measurements are considered within this work: ultraviolet index, luminosity, color temperature, red light, green light, blue light, and clear light. ANNs are trained, validated, and tested using all combinations of these different light measurements as inputs. The classification accuracy of each of these variations is compared and used to determine the effectiveness of the individual measurements for classification purposes. The results of this work revealed that the color temperature measurement was particularly effective for detecting outdoor exposure when used in conjunction with at least one other measurement type. Additionally, it was found that the ultraviolet index may not be a necessary component for classification algorithms.