Not using AKD1000 yet but it is another proven use case:

Vision and Deep Learning-Based Algorithms to Detect and Quantify Crac
ks on Concrete Surfaces from UAV Videos

by Sutanu Bhowmick ¹,Satish Nagarajaiah ^1,2,* andAshok Veeraraghavan ³

¹

Department of Civil and Environmental Engineering, Rice University, 6100 Main Street, Houston, TX 77005, USA

²

Department of Mechanical Engineering, Rice University, 6100 Main Street, Houston, TX 77005, USA

³

Department of Electrical and Computer Engineering, Rice University, 6100 Main Street, Houston, TX 77005, USA

^*

Author to whom correspondence should be addressed.

Sensors 2020, 20(21), 6299; https://doi.org/10.3390/s20216299

Received: 28 September 2020 / Revised: 26 October 2020 / Accepted: 30 October 2020 / Published: 5 November 2020

(This article belongs to the Special Issue Sensors, Big Data Analytics and Modeling for Infrastructure Monitoring and Maintenance)

Download PDF

Browse Figures

Abstract

Immediate assessment of structural integrity of important civil infrastructures, like bridges, hospitals, or dams, is of utmost importance after natural disasters. Currently, inspection is performed manually by engineers who look for local damages and their extent on significant locations of the structure to understand its implication on its global stability. However, the whole process is time-consuming and prone to human errors. Due to their size and extent, some regions of civil structures are hard to gain access for manual inspection. In such situations, a vision-based system of Unmanned Aerial Vehicles (UAVs) programmed with Artificial Intelligence algorithms may be an effective alternative to carry out a health assessment of civil infrastructures in a timely manner. This paper proposes a framework of achieving the above-mentioned goal using computer vision and deep learning algorithms for detection of cracks on the concrete surface from its image by carrying out image segmentation of pixels, i.e., classification of pixels in an image of the concrete surface and whether it belongs to cracks or not. The image segmentation or dense pixel level classification is carried out using a deep neural network architecture named U-Net. Further, morphological operations on the segmented images result in dense measurements of crack geometry, like length, width, area, and crack orientation for individual cracks present in the image. The efficacy and robustness of the proposed method as a viable real-life application was validated by carrying out a laboratory experiment of a four-point bending test on an 8-foot-long concrete beam of which the video is recorded using a camera mounted on a UAV-based, as well as a still ground-based, video camera. Detection, quantification, and localization of damage on a civil infrastructure using the proposed framework can directly be used in the prognosis of the structure’s ability to withstand service loads.

Keywords: computer vision; morphological operations; unmanned aerial vehicle; U-Net

1. Introduction

Condition assessment of civil engineering structures for its safety and remaining lifetime has been in focus for the past couple of decades. Mostly, it consisted of harnessing dynamic response by attaching acceleration and displacement sensors with further post-processing of the data to evaluate the presence of damage in those structures. This method provides the global behavioral pattern of the structure which may sometimes provide local damage indications depending on the kind of structure and the spatial density of the sensors.

On the other hand, local assessment of structural damage is mostly carried out by visual inspection by experts. Otherwise, by placing contact-based strain sensors or by carrying out acoustic-based non-destructive testing. Sun et al. [1] present a detailed review of structural health monitoring methods based on the big data and artificial intelligence algorithms for bridges, as well as identifies challenges associated with them. However, given the size of the structure itself, the whole process becomes time-consuming and prone to human errors. In addition, in some regions of the structure, it is difficult for human beings to gain access. Contact-based sensors also suffer from high maintenance due to the wrath of the exterior environment they are exposed to.

Recent advances in the application of non-contact camera-based structural health monitoring got huge incentive with the development in high-resolution cameras coupled with robust computer vision algorithms. Some of the real-life applications of such computer vision algorithms include face detection in mobile cameras, motion detection for surveillance, traffic sign, and pedestrian detection in autonomous cars. Similar technologies can be customized to detect local damages on structural surfaces using video measurements. Video acquired from surveillance cameras installed on important structures, like bridges, is a valuable source of data for damage detection. Moreover, fast and automatic damage inspection of large scale structures can be performed using the unmanned aerial vehicle (UAV) equipped with a digital camera and onboard microprocessor. Koch et al. [2] present a review of computer vision algorithms that have been used for damage detection and condition assessment of civil infrastructure.