Utilising the power of AI for Flood Detection on Queensland Roads

Abstract

Councils typically enter subscription service agreements with camera providers to monitor roads. In 2022, a number of these camera providers folded, leaving councils like Carpentaria Shire Council with camera hardware that no longer functioned due to the camera feeds requiring specific systems only offered by the abandoned company. Carpentaria Shire Council saw an opportunity to self-manage their camera network to future-proof their management of flooding over roads, reducing reliance on external providers. In collaboration with council, the LGAQ Lab has delivered a low-cost innovative solution for assessing flooding over roads that used artificial intelligence (AI) to undertake the detection of flooding over roads, exclusively from image reads. The LGAQ flood detection AI model can detect flooding over roads with 98.6% accuracy using only visual feeds and no bespoke hardware locked into one provider. This significantly lowers operating costs for councils and provides greater flexibility for their asset management practices. Images are scanned every 15 minutes, and once flooding is detected, the council officer is automatically notified by email. This eliminates manual checking of camera feeds (as done with traditional CCTV) and the need for in-person inspections, returning entire days of productivity to the council and increasing the safety of council officers. There are a range of flood detection providers in the market, but councils have flagged that they are expensive, require long-term service agreements and often lead to unsupported hardware when the company exits the market. That is why the LGAQ took a hardware agnostic approach to allow councils to adopt the AI detection method on camera networks self-managed by the council. This eliminated long-term risks, improved asset management capabilities, and increased council process efficiencies through automating flood detection in the cloud. In this presentation Matt Oxlade will share key insights into how an initial problem statement grew legs to become a project that will maximise the efficiency of existing council resources and improve outcomes for local communities.