Improving inspection processes through aerial surveillance
Aviation & Aerospace, Drones, Aerial Surveillance, Software
CentrovisioN is a leading Unmanned Aerial System (UAS) company headquartered in Bangkok, Thailand. CentrovisioN UAS can be leveraged to carry out tasks such as capturing aerial survey images, performing building facade inspection and generating high resolution aerial panoramic images. The company is focused on using advanced technology, such as computer vision and machine learning, to improve the efficiency and accuracy of building inspection processes.
One key area of focus for CentrovisioN is the inspection of building facades. Traditional methods of inspection, such as manual inspection or reliance on visual inspection alone, can be time-consuming, costly, and prone to errors.
To address these challenges, CentrovisioN has partnered with Datature, a leading MLOps platform for computer vision technology, to develop advanced deep learning models for building facade inspection. These models are trained to identify and classify defects and abnormalities on building facades, such as sealant gaps, clips, missing tiles, discoloration and other damages. The models are integrated into CenInspect to automate defect inspection and increase inspection throughput by 10x.
In order to train a MaskRCNN model, the images captured by the drone need to be first annotated and then augmented. Varying weather conditions could affect the accuracy of the model drastically; therefore, proper image augmentation techniques can help generate more robust models. Once the model is deployed, Active Learning is implemented so that the model can receive human-in-the-loop feedback to iteratively improve its accuracy. Finally, the team also leverages Orthomosaic and Photogrammetry to stitch the images together.
By using these computer vision models, CentrovisioN is able to significantly improve the efficiency and accuracy of building facade inspections. The models are able to analyze building facades in real-time or in batches at a high throughput, identifying defects and abnormalities that may have been missed by manual inspections. This helps to reduce inspection time and costs, while also improving the accuracy and reliability of the inspection process.
Overall, CentrovisioN's use of Datature's computer vision models has helped to transform the building inspection process in Thailand. By using advanced technology to analyze building facades, the company has been able to significantly improve the efficiency, accuracy, and safety of the inspection process. As a result, CentrovisioN has become a leading provider of building inspection solutions in Thailand, and is well positioned to continue driving innovation in this space.
Datature’s user-friendly interface has expedited building Computer Vision models, which are then used by CenInspect to carry out defect identification and analysis in a more streamlined and efficient manner - Thamma Varangoon | Founder at CentrovisioN
Datature is proud to support CentrovisioN’s mission to improve inspection processes through aerial surveillance and we are excited to see what’s next for CentrovisioN.
If you are an innovator looking to unlock the potential of deep learning to transform the UAS industry, get in touch with us!
CentrovisioN was established in 1991 with its original objectives to be a leading business and management consultant. In 2010, CentrovisioN was engaged in a large-scale aerial survey of plantations in Indonesia, and that was the start of our UAV development. In the first year of the survey operation, CentrovisioN was struggling in finding the right solution for the suitable and workable equipment to finish the tasks. Many off-the-shelf commercial UAV products were tested but found to either not to pass the stringent requirements or lack flexibility to cope with different tasks, therefore CentrovisioN commenced its own research and development to come up with Unmanned Aerial System (UAS), which comprises their own design airframe, autopilot control, radio and data link communication, ground station as well as image processing software and hardware. CentrovisioN UAS can be adapted to meet all tasks in all working environment, thus providing solutions to all remote sensing and aerial survey requirements.
Datature is an end-to-end MLOps platform that allows teams and enterprises to build computer vision models without a single line of code. Teams can manage datasets, annotate, generate synthetic data, train and deploy - all in a single, secure cloud-based platform. With the rise of citizen data scientists, deep tech companies, and enterprises looking to adopt deep-learning - Datature equips these startups / experts with the tools required to build their own capabilities easily within weeks.