AI x Drone utilized Crack Detection System

Benefits of automating

  • One of the main benefits of automating the crack detection process is the  use of drone cameras in towers, buildings, bridges.Another advantage is the significant cost savings that allow for frequent inspection cycles. Citizens can lead a safer life by conducting frequent inspections.

・Infrastructure such as bridges and dams are prone to deterioration with heavy use. This unavoidable process is causing urgent maintenance issues and restore the problems by quick findings.

・system can also be customized to develop inspection tools that correspond in each objects and new constructions.


Until recently, we took a picture of the building with and judged it with the human eyes visually  without using the drone, or the inspecting by local engineer that leads to judge that the bridges and buildings had deteriorated and made Decision by human eye that repair is a must.Not only does this work take time, but also rely on visual inspection. It has been   we introduced is judgments NOT depending on the workers’ at the site. The service we are proposing is the service is AI equipped drone, and work efficiency can be dramatically improved and inspections can be carried out according to certain criteria that is come from data base .Our image processing technology enables high-speed and high accuracy processing by bundling Edge AI with on-site PC. Novius, a leading AI imaging company. As a result of them, it enables real-time (high-speed processing) of images utilizing by AI.

AI can process and analyze images taken by drone at site

AI drone detects a crack and determines its length and width. Thru deep learning as the detection method, it is possible to detect accurately with AI without erroneous detection due to dirt.

Our AI is basically processed by the edge, but if there is not enough information contained in the device at site, data that needs to fulfill AI can transfer from cloud and the image is extracted from the enormous amount of data to speedily detect cracks at the site. so analysis can be done there.

The characteristics of our imaging processing is that the crack and damage detection is separated into two categories by performing the cloud server processing in two divisions after the “Edge AI processing” on the PC. With AI edge processing, it is possible to immediately check for cracks and damaged areas on site. It also reviews the flight route of the drone such as the part you want to shoot again.

How to Utilization

・ This AI Drone works on the combination of AI (artificial intelligence), machine learning, data analysis and communication technology.

Depending on the needs of consumers, this system can be easily customized and depends entirely on the expected data.

Surface cracks in  building structures, towers, welds, dams, wind turbines are the detection that will be available in this technology.

Advances in image processing technology can also determine the crack direction. (Horizontal, Vertical, or Ramp type).

By enhancing the system of AI Drone, it is possible to quantify crack size.

Actual effects of on-site introduction

・One of the serious problems World face is the increase of imminent big earthquake and meteorological disaster like heavy rain. This AI system can usefully detect cracks caused by these natural disasters.・The aging of social infrastructure such as highway bridges, sewers and roads is one of the major fears and worries. In particular, cracks in highway bridges and general road bridges can hinder not only human life but also economic activity. 
・To ensure the safety, Introduction of AI drone detection is a MUST.・This AI drone also helps overcome structural problems in building and social infrastructure.
・AI utilizing this AI technology for crack detection can reduce labor concerns and high cost challenges, and increase productivity.

On-Site Detection Work

A drone is used to shoot the entire structure and collect image data. It is used for grasping the entire structure and mapping of damaged parts. Take  picture of the details and based on the image. Novius AI engine detects the damaged areas and outputs the result of analysis at site.

Although the final confirmation is done by human, it can detect where the damage is in real time, supporting the decision and analysis at site

It is suitable for inspection of high rises and dangerous places because it is aerial shot by drone, and it not only solves the labor shortage, but also helps ensure the safety of workers.

※Hardware Features※

①Edge PC used-

PC used for Edge AI requires the recommended specifications to operate AI efficiently.

NVIDIA GTX -1650 or higher, Intel/AMD CPU i5/i7, Memory of 8G or more

* High-speed image processing cannot be performed without a graphic board.

Customers who use drones can select cameras that can detect cracks and 

damages The captured image can be detected and analyzed. The OS runs on Windows 10.AI works just by downloading the Edge AI drone software to your devices

②Server use/When using a server, it works regardless of the device performance. smart phone, 

      Tablet, Normal PC. Images can be detected and analyzed on those devices. The industry's first system that uses server technology