About 7 percent of the world's labor force is employed in the construction industry, and about 10 trillion dollars are spent per year on construction-related activities. By the time of 2050, the construction industry is poised to serve a predominantly 9 billion urban population. Despite the expanding market, the construction industry is found to be one of the least digitized industries according to McKinsey Global Institute (2017). Adopting the latest technology can be daunting, but Artificial Intelligence and Machine Learning that have made an impact in other industries are beginning to emerge in the construction industry. Among a wide spectrum of AI technology, computer vision is a natural fit into the construction industry.<\p>
During my first undergraduate degree, I volunteered in a teaching program for an underrepresented Thai community in Hong Kong and worked with children who are unable to pursue higher education due to their family’s economic status. Despite this, all the children I worked with had a strong desire to learn. For the first time, I realized how fortunate I am to enjoy the privilege of a university education. Knowing that there are many more underserved populations around the world, a sense of equality compelled me to do more to make a positive impact in the lives of others.
Productivity and efficiency are crucial in the construction industry, and a change that speeds up the process can mean millions of dollars saved or weeks of work time spared from unexpected delays. One such change is the adoption of on-site cameras coupled with a real-time object detection algorithm backed by deep neural networks. On one hand, cameras can be mounted on workers, robots or drones which can collect thousands of site images and video data. An object detector can then be trained and deployed to detect defects such as cracks, leakage, poor workmanship, and deviation from standardized work plans. On the other hand, the computer vision system can monitor various construction activities such as movements of equipment, progresses of tasks, and interactions of workers in a real-time manner. Such information can then be compared against Building Information Modeling (BIM) specifications and transformed into valuable insights to track the project progress and optimizes the workflow.<\p>
Another application of computer vision lies in the automation of repeated work. Companies are using self-driving construction machinery to perform repetitive tasks such as pouring concrete, demolition welding, and bricklaying. In semi-autonomous scenarios, machinery such as tower cranes and bulldozers equipped with off-the-shelf hardware (e.g. cameras) are able to calculate the optimal movement and guide the machine operator to the most efficient trajectories. In short, computer vision systems turn heavy machineries into smart and autonomous robots that are more accurate and efficient than their human counterparts. <\p>
In a foreseeable future, AI technology will automate many other construction tasks in unprecedented ways. Although AI is unlikely to replace the human workforce, the business models in the construction industry will be altered, and, consequently, a more balanced project management scope triangle measuring time, cost and quality can be achieved.<\p>