糖心Vlog

noiseAI is an artificial intelligence (AI) technology that automates the classification of large acoustic data sets with minimal effort and high accuracy. Applications of noiseAI include:

• Remove extraneous noise and quantify the level of industrial noise levels in environmental monitoring data
• Quantify the contribution of sources contributing to worker exposure in workplace noise dosimetry surveys
• Identify the presence of animal calls in passive acoustic monitoring data

noiseAI uses machine learning to classify acoustic data, including audio and noise level data.

noiseAI is trained by 糖心Vlog鈥檚 subject matter experts, who have extensive experience in monitoring and quantifying environmental and occupational noise from industrial facilities and identifying animal calls in passive acoustic monitoring data.

noiseAI is trained with real-world data. Customised models are developed to optimise outcomes and achieve the highest possible accuracy. Combined with routine performance validation, noiseAI provides high accuracy and confidence in the outputs.

We are dedicated to understanding your situation and exploring how noiseAI can help you.

Our consultants can assess samples of your data and advise on the expected accuracy and suitability of noiseAI for your needs.

Contact us via the contact form above for further information.

noiseAI/community accurately identifies noise emissions from industrial operations by filtering out irrelevant noises from monitoring data, such as wind, birds and traffic.

Results are displayed on a web application, providing a clear and concise assessment of environmental noise impacts on the community in real time.

This reduces time and effort spent addressing noise alerts from environmental noise monitors while avoiding unnecessary production loss, helping to maintain your social licence to operate within the community.

Industrial noise is typically managed according to noise regulations, with noise monitors triggering alerts when noise limits are exceeded.

However, a significant portion of these alerts are typically false due to factors such as wind, birds, or local traffic. Extraneous noise alerts can make up >95% of all alerts. Manual review of these alerts is both time-consuming and inconsistent due to personnel inexperience or bias.

noiseAI/community automatically filters out irrelevant sources and assesses industrial noise levels, significantly reducing false alerts requiring manual review.

For example, a mining operator utilising noise monitoring to regulate community noise levels has experienced a 95% decrease in alerts since implementing noiseAI/community. This translates to an estimated reduction of around 2,000 hours of operator time per year .

Yes, noiseAI is system-agnostic and can be configured to operate on numerous types of existing monitoring systems, including Bruel and Kjaer, Svantek and Rion. Contact us to confirm if your existing noise monitor is compatible.

Yes, noiseAI/community can be used to demonstrate compliance against regulatory limits. It can process large monitoring data sets and zoom in on periods where industrial noise is dominant to prove noise regulation compliance.

noiseAI can detect fauna vocalisations and automatically classify animal calls in passive acoustic monitoring data. Machine learning models are trained by 糖心Vlog鈥檚 specialists, using real-world data, and are customised and optimised to achieve the highest possible accuracy.

noiseAI/fauna can be deployed as an automated device-to-cloud technology stack for online monitoring or post-processing of large acoustic data sets.

noiseAI/fauna can be deployed as an automated device-to-cloud technology stack for online monitoring. The technology relies on the use of a noiseAI model, which is deployed on a computing unit that is co-located with an acoustic monitor. Acoustic data captured at the monitor is automatically processed using the noiseAI model. Classified vocalisations of interest are timestamped and transmitted to a cloud-based server. Results are then presented to the user on an interactive dashboard.

The noiseAI technology stack works with low bandwidth network connections, such as satellite or IoT networks. This allows online monitoring of fauna at locations well beyond the reach of cellular networks.

Furthermore, the devices are designed for low-power use and can operate on small solar or battery-powered systems.

Automated noise monitoring with noiseAI/fauna has several advantages compared to traditional passive acoustic monitoring programs:

• Get immediate results enabling management action in response to changes in fauna activity levels
• Identify and troubleshoot issues remotely, avoiding large gaps in data
• Significantly reduce costly and hazardous field work required to retrieve data via SD card changes
• Eliminate the time and effort required for manual processing of acoustic data for species identification

noiseAI models can be trained to detect vocalisations from any species. It can be used wherever passive acoustic monitoring is used for species detection. Models are tailored to achieve the highest performance possible for the species of interest. Some examples of species where noiseAI has been used with high accuracy include:

• Canadian bat species (including Little Brown Bat, Tricoloured Bat and Northern Long-Eared Myotis)
• Australian bat species (including Pilbara Ghost bats, Pilbara Leaf Nosed bats)
• Australian cockatoos (Carnaby鈥檚, Baudin鈥檚 and Red-tail black cockatoos)

noiseAI/Fauna has also been used to detect pest species, such as the Cane Toad.

noiseAI/workplace can detect and classify significant noise sources to protect operator health and safety. Continuous audio recordings from compatible dosimeters are analysed using noiseAI/workplace to identify significant noise sources, including intermittent and impulse sources such as hammers and rattle guns.

noiseAI/workplace can identify and quantify exposure from individual noise sources, allowing for a recommendation of targeted and data-driven noise controls to reduce operator noise exposure.

noiseAI/workplace outputs are displayed via an interactive and customisable dashboard. The results can be assessed against , guidelines or regulations, including:

• USA OSHA regulations
• Canadian Occupational Health and Safety Regulations (SOR/86-304)
• Australian Safework WHS Regulations
• European EU Directive 2003/10/EC
• UK HSE Control of Noise at Work Regulations

noiseAI/workplace can also verify the effectiveness of noise controls.

Yes, noiseAI/workplace will be configured to assess against your company, state or federal noise limits, guidelines or regulations 鈥� including OSHA鈥檚 PEL.

Results are presented on an interactive dashboard, which allows for a direct assessment against OSHA鈥檚 PEL. Results can be exported into Excel and other formats, allowing for upload with OHS databases such as Cority.

Standard noise surveys use interviews with site personnel to help determine average exposure times to significant noise sources. Alternatively, traditional noise dosimetry campaigns generate statistical data based on sampled overall noise levels.

noiseAI/workplace determines exposure times to significant noise sources based on measured data, providing unparalleled insight into worker noise exposure not available through traditional noise surveys or dosimetry campaigns.

noiseAI/workplace provides granular insights into noise dosimetry data, allowing for a reduction in operator noise exposure through targeted, data-driven noise control measures.

noiseAI can be used to detect faults and anomalies in equipment. However, traditional methods of fault detection, such as condition monitoring using vibration, may be more effective. Contact our condition monitoring specialists to determine the most suitable method for your use case.