An image augmentation method for detecting construction resources using convolutional neural network and UAV images

Images acquired by UAV can be analyzed for resource management on construction sites. However, analyzing the construction site images acquired by UAV is difficult due to the characteristics of UAV images and construction site images. This paper proposes an image augmentation method to improve the performance of an object detection model for construction site images acquired by UAV. The method consists of three techniques: intensity variation, image smoothing, and scale transformation. Experimental results show that the method can improve the performance of the detection model (Faster R-CNN) by achieving a recall and a precision of 53.08% and 66.76%, respectively. With future studies, the method is expected to contribute to UAV-based resource management on construction sites.