A solution for reducing electricity in residential sector using image processing

Abstract

Energy saving is a critical issue that should be addressed in a worldwide scale. In the residential sector of Sri Lanka, there are many houses. Each household on average includes four people and has diverse electronic needs to be fulfilled. This paper proposes a solution to reduce the electricity consumption of residential sector. The solution has the ability to manage the use of electricity consumption of households. It identifies each and every household electric item and connects through Wi-Fi. Each household electric item which has the ability to connect to a Wi-Fi network, will be connected to the system via the routers port forwarding function. The user has the ability to check the system and identify which electric item is wasting energy and then the user can switch it off remotely through the system. Furthermore, the proposed solution is equipped with image processing algorithms. Image processing is fast, flexible and opens a whole new world of real time computer vision. A video camera located in several places in the house is used to identify presence of humans and then automatically switch off unnecessary electronic items. The proposed detection process depends on the light condition, camera angle and the efficiency of the real time detection. Matlab’s SVM classifier people detection algorithm was used as the image processing algorithm. One thousand six hundred images were split equally into two data sets as images with humans, and images without humans. The analysis revealed a unique threshold value as 6 220 800 in images to identify humans images in it. In the future, the system is envisaged to connect to an IoT (Internet of Things) platform to derive more benefits to the end user.