Application for Roof Type Analysis and Component Counting from Roof Plan Image

In this paper, an application for analyzing roof types and counting components from roof plan images using Python OpenCV was proposed. The goal of this study is to help users to estimate roof construction costs. The application can determine the type of roof: Hip, Gable, and Dutch; also, predict and count the number of crucial components: Hip valley, Raft valley. Our proposed rule-based technique was used to utilize sub-image in particular coordinates. Regarding the experimental setup, variation of roof plan types as Easy, Normal, and Hard (depending on the number of roof types and sub roof in a single house) were tested. As a result, the accuracy of our proposed algorithm was nearly 100% in Easy and Normal types of images, which is accurate and acceptable to be used for estimating in cost calculation.