logo

AI-powered Road Analytics in Multiresolution Aerial Images

Accurate, detailed and up-to-date road networks are critical in many applications, autonomous driving being one of them. Currently, the existing road network data lacks sufficient detail and currency necessary to support various use cases such as asset management and monitoring as well as smart transport networks.

To address this issue, we developed innovative transfer and deep learning models to detect road furniture and safety features in both 10cm aerial and satellite images with a similar quality.

Get the white paper

Optimising pest control with artificial intelligence

Snails consume crops and contaminate grain, leading to reduced harvest quality and subsequent financial loss for growers. Current methods for snail control are inefficient and expensive. To address this challenge, we developed an artificial intelligence solution to identify snails in smartphone images and create density map of snails across the farm. This critical insight was used to implement variable rate technology and optimise the rate of chemical bait application. Through our technology, growers were able to minimise the cost associated with pesticides and reduce negative environmental impacts due to overuse of chemicals.

Get the white paper

AI-powered fish survey

Marine ecology research utilises underwater cameras for wide-ranging applications. The enormous data being gathered from underwater videos needs to be consumed in an effective way to automatically detect and classify the species and maintain their count for estimating the biodiversity and protecting endangered fish species. In this project, we developed a set of AI technologies and computer vision processing workflows for fish taxonomy classification and biomass estimation in underwater stereo videos.

Mine rehabilitation assessment through machine learning in pleiades images

Monitoring using remote sensing techniques is increasingly playing a role in the assessment of mine-site rehabilitation. In the past, the major impediment to the uptake of technology has been the lack of resolution and inaccurate image understanding techniques. Recent advances in sensor technology and the rise of Artificial Intelligence (AI) have opened up new possibilities for mining industry.

Get the white paper

Deep learning and pleiades images for sandalwood forest inventory generation

Forest inventory is critical for effective management of plantation farms. A detailed, accurate and up-to-date inventory can help forestry companies make informed decisions and improve efficiency. Inventory components such as trees stocking, weed infestation and gaps between trees could be measured in direct way through sampling or via remote sensing. In general, direct methods are very labour intensive and costly, and subject to sampling error.

Get the white paper

Deep learning for building footprint change detection

A number of industries need an up-to-date and accurate building footprints and change map. In real estate, it empowers property professionals and real estate agents to check for any potential adverse possession claims and compliance with approved building plans. In Insurance industry, insurers need to assess damages to buildings in disaster events such bushfire and flooding which requires accurate change detection map.

Get the white paper