There have been many studies examining computer vision-enabled research in construction. They use an array of algorithms to recognize ‘objects’ in images and videos to classify, segment or verify information based on generic databases. The emphasis has generally been placed on measuring the accuracy of object detectors, such as Faster Recurrent-Convolution Neural Networks and variants thereof. However, limited, if any, attention has focused on determining the benefits of such technology during construction beyond the broad evidence that it can help detect problems quickly and accurately. This article brings to the fore the need for researchers in construction to examine the benefits realization management of computer vision-enabled applications to ensure value is added to an organization’s business operations.

有许多研究探讨了建筑中计算机视觉支持的研究。他们使用一系列算法来识别图像和视频中的“对象”，以基于通用数据库对信息进行分类、分段或验证。重点通常放在测量物体检测器的准确性上，例如更快的循环卷积神经网络及其变体。然而，除了广泛的证据表明这种技术可以帮助快速、准确地检测问题之外，人们的注意力（如果有的话）也有限地集中在确定这种技术在施工过程中的好处上。本文凸显了建筑研究人员需要检查计算机视觉应用程序的收益实现管理，以确保为组织的业务运营增加价值。