JoWUA

Abstract

Many commonplace issues may be addressed with the help of the Internet of Things (IoTs). We provide an efficient solution to test classification and status of flowers images by combining Deep Learning (DL) and IoTs. Due to the diversity of flower species, classifying and determining the status of flowers is a difficult undertaking. in this study, two-phase DL to differentiating between flowers of several species are developed into the application of mobile that discovers the type of flower so that describing the amount of water required for the daily irrigation for every kind of flower. First, the flower region is automatically segmented using Markov Random Fields (MRF) approach. Second, build a robust DL by using four models such as VCG-16, inception-V3, MobileNet-V2, and RestNet-18 to distinguish the different flower types. In this way, the system keeps track of the amount of water in the soil and uses water wisely while appreciating the user. To evaluate the flower's state, which is being tracked by nearby sensor devices, a cloud-based server and a mobile application are used in the flower status tracker implementation. The results show that the VCG with segmentation has the highest accuracy than other models but it has longer time in implementing.