Computer Vision Based Empty Floor Space Recognition System For Infrastructure Management

Efficiently monitoring occupancy to optimize space utilization is a fundamental aspect of infrastructure management that can lead to better resource allocation and reduced costs. This research aims to create a system that can identify vacant floor space, which is essential for managing various aspects such as facilities, real estate, retail, logistics, warehouses, and smart city infrastructure. For this, the detection of on-floor object should be done properly. To ensure that the system is correct, the images from the dataset must be properly pre-processed. The paper gives comprehensive overview about system which performs empty floor detection using different machine learning classifiers. The research done is basically focused on machine learning model which can give more accurate results. To evaluate the precision of the system, different machine learning classifiers such as Decision Tree, K Nearest Neighbour (KNN) and Random Forest are applied. Systems uses Binary Robust Invariant Scalable Key points (BRISK) feature descriptor to obtain the key points which helps all the classifiers to provide high accuracy. Different classifiers offered varying degrees of accuracy among those Random Forest classifier has highest accuracy of 94.67%.