By Dr Ehsan Noroozinejad, Senior Research Fellow at Urban Transformations Research Centre (UTRC)
Infrastructure monitoring is crucial in ensuring critical structures’ safety and efficient functioning. This is especially true in Australia, a country known for its diverse and challenging environment, which requires robust monitoring systems to address the unique infrastructure needs of the country.
In recent years the construction industry has experienced a significant transformation, with advancements in Artificial Intelligence (AI) and Machine Learning (ML) promoting innovation in the sector. This article will explore the potential of drone-based autonomous inspection and monitoring, empowered by AI, to enhance infrastructure monitoring in Australia.
In recent years, various researchers and civil engineering applications have successfully utilised convolutional neural networks (CNNs) for image classification and object detection. Metal surface defects detection, post-disaster collapse classification, joint damage detection, concrete crack detection, pavement crack detection, and structural damage detection are just a few examples of the wide range of applications.
These studies have demonstrated the effectiveness of CNNs in analysing visual data and extracting meaningful information. Object detection, which involves classifying and localising objects, has also seen significant advancements with the introduction of CNNs.
Traditional methods like the histogram of oriented gradients have been replaced by region-based CNNs (R-CNNs). These methods utilise region proposal functions to localise and segment objects, improving performance and accuracy compared to previous approaches.
Faster R CNN is a notable advancement that further enhances speed and accuracy in object detection. In the field of civil engineering, region-based CNN methods have been successfully implemented. Studies have used Faster R-CNN to detect structural damage types, including steel delamination, corrosion, and concrete cracks.
The efficiency and effectiveness of Faster R-CNN have been demonstrated in various scenarios, such as shield tunnel lining defect detection and concrete defect detection with geolocation.
The role of AI and ML in the construction sector
AI and ML technologies have revolutionised the construction industry, increasing project completion efficiency and safety. Real-time data analysis provided by AI helps in construction planning, enabling informed decisionmaking and risk mitigation. By optimising processes, reducing errors, and improving project timelines, the integration of AI in construction operations has proven highly beneficial.
One emerging field – robotics construction – leverages AI to enhance productivity and safety by deploying autonomous construction vehicles and robotic systems capable of performing complex tasks with precision.
Autonomous monitoring of infrastructure
An area where AI has made remarkable strides is in the autonomous monitoring of infrastructure. Traditional inspection and monitoring methods are often time-consuming, expensive, and pose risks to personnel.
However, the utilisation of drones equipped with sophisticated sensors and AI algorithms presents a revolutionary solution. In Australia, with its vast landscapes and remote areas, drone-based autonomous inspection and monitoring systems offer an efficient and cost-effective approach to infrastructure management.
Advantages of drone-based autonomous inspection and monitoring
Drone-based inspection and monitoring systems powered by AI bring numerous advantages to infrastructure management in Australia.
Drones possess the capability to access hard-to-reach areas, such as bridges, pipelines, and power transmission lines, without compromising the safety of human inspectors.
Equipped with high-resolution cameras and LiDAR sensors, drones capture accurate and detailed data, which is then analysed using AI algorithms. This data enables the identification of potential structural issues, such as cracks, corrosion, or deformations, with greater accuracy and efficiency.
Moreover, drones facilitate rapid data collection over large areas, reducing the time required for comprehensive inspections. Real-time data processing provides timely insights, enabling proactive maintenance and repair strategies.
By integrating ML algorithms, the system can learn from past inspections, enhancing its ability to identify anomalies and predict maintenance needs, ultimately leading to more efficient infrastructure management.
AI and carbon neutrality: accelerating decarbonisation efforts
AI-powered drone-based autonomous inspection and monitoring systems have the potential to accelerate decarbonisation in Australia’s built environment. Energy inefficiencies can be identified and retrofitted by leveraging thermal imaging and AI algorithms to reduce emissions.
Drones can also assess the suitability of renewable energy infrastructure and optimise construction processes to minimise waste and energy consumption—AI’s ability to analyse data on materials’ environmental impact aids in selecting sustainable options.
Additionally, drones can monitor green infrastructure and provide real-time emission monitoring, supporting the transition to a carbon neutral future. Integrating AI into infrastructure monitoring empowers evidence-based decision-making and drives the adoption of sustainable practices. Australia can leverage AI technologies to enhance infrastructure monitoring while advancing its carbon neutrality goals.
Circular economy and AI-enabled design processes
Australia’s commitment to a circular economy, aimed at eliminating waste and promoting sustainable resource use, aligns perfectly with the integration of AI in infrastructure monitoring. AI can accelerate the development of circular economy principles by enhancing the design and development of products, components, and materials.
Through iterative machine-learningassisted design processes, AI enables rapid prototyping and testing, facilitating the creation of environmentally friendly and economically viable infrastructure solutions.
What does the future hold?
Drone-based autonomous inspection and monitoring systems, bolstered by AI, hold immense promise for enhancing infrastructure monitoring in Australia.
By harnessing the power of AI and ML, Australia can benefit from efficient and cost-effective monitoring of critical structures. These technologies offer timely and accurate data, enabling proactive maintenance strategies and supporting the country’s commitment to a circular economy.
As Australia prioritises the safety and sustainability of its infrastructure, embracing the potential of AI-driven monitoring systems is a crucial step towards building a resilient and futureready environment. While ANN-based classification has shown reasonable accuracy in civil engineering applications, more training data is needed.
Proper pre-processing of training data is crucial to identify localised damage accurately. Images often contain multiple damage states, and it is essential to pre-process the data to differentiate between damaged and undamaged areas effectively.
In conclusion, AI and ML technologies have proven to be powerful tools in civil and infrastructure engineering, enabling accurate image classification and object detection. Integrating these technologies with drone-based autonomous inspection and monitoring offers tremendous potential for enhancing infrastructure monitoring in Australia.
By harnessing the capabilities of CNNs and advanced AI algorithms, Australia can achieve more efficient and costeffective infrastructure management, ensuring the safety and longevity of critical structures.