Sensor Analytics Australia is a research and development enterprise; specialising in systems and software for vibration, vision, and sound sensors in IoT and robotics. SAA provision data collectors (DACs) for datamining, pattern recognition, and visualisation of big data sets and streaming data sources. SAA also implement fully integrated end-to-end systems for Internet-of-Things (IoT) in collaboration with highly experienced IoT equipment and systems providers.
SAA provisions Machine Learning (ML) and Optimisation modelling, Artificial Intelligence (AI) - trained neural networks, and computer vision (CV) for industrial application.
-BigIoT™
S-BigIoT™ is an open source Raspbian image for RPI3 and above. S-BigIoT collects inputs from a variety of directly attached sensors, temperature, vibration, moisture etc supported IoT motes, ThingWave,NCD and USB devices. mics, cameras, Arduino boards It securely and reliably uploads device data to an S-BigAnalytics instance running anywhere on the Internet. Just-in-time (time_t=now) reports can be obtained on key operational statistics and business intelligence. Equipment monitoring, production area statistics, historical analyses, OHS related data logging, data forensics, supply-chain forecasting, predictive maintenance, perimeter security, public area statistics, computer-vision, heat-maps, alarms, quarantine, biosecurity, negative-pressure monitoring, chemical-biological-radiological (CBR) dispersion, environmental monitoring, flood-mapping enhancement are some of the functions provisioned by this system. Realtime sensor readings are also available through a live web-based dashboard. S-BigAnalytics and S-BigIoT together form a unified end-to-end IoT data acquisition & processing framework for industrial, workplace, public health, security, and environmental applications.
-BigAnalytics™
S-BigAnalytics™ is an open source big data, machine learning, and datamining framework for IoT.
It is supported on Linux, MacOS, and MS-Windows. S-BigAnalytics™ is also provisioned as a ready-to-use SaaS on Lateral Blast Cloud. S-BigAnalytics™ is a virtual machine based platform that can be installed on any personal computer or in the cloud. It processes raw sensor data such as thermal, audio, and image/video, into:
- Events
- Business Intelligence (BI) about such events e.g. linked events
- Insights into actions
- Detecting anomalies and outliers
- Gathering use-patterns and statistics
- Predicting important events
S-BigAnalytics™ can apply its machine learning models to millions of records for gathering business intelligence in real-time.
S-BigAnalytics™ provisions a framework for IoT data acquisition and hosting machine learning and datamining apps.
ystem Information
SAA provision an essential component for IoT systems*; in the form of a platform that handles real-time data collection, visualisation, and alerts. This platform can be deployed for a range of IoT projects such as but not limited to smart city and smart industry applications - manufacturing equipment, warehouses, civic buildings, public spaces, perimeter security, early warning, hospitals etc.
SAA data collectors (DACs) achieve the highest economy of scale with the entire data collection and visual analytics system being integrated within a single compact platform, making online data analytics and datamining operations possible for dynamic data streams and static data sets simultaneously. In doing so we can provision a very high level of performance at an affordable cost to the end user. The platform scales up with each new DAC, bringing additional storage and processing capacity.
SAA provide a complete in-house solution, with easy to understand documentation, for customers to self-manage their data collection and analytics**.
SAA systems are fully self-contained; no worries about data ownership or to incur on-going and ever increasing cloud rental charges.
If requested, SAA can also provide end-to-end support for the installations. We provide full system analysis and project management services for large projects e.g. multi-site or highly specific deployments, working closely to achieve satisfactory outcomes for all stake holder.
- *Please contact us with the details of your IoTs to check if these can be supported by the DAC.
- **SAA Analytics Levels
- Bir: Charts and Dashboards
- Iki: BI Questions, Datamining
- Uch: Machine Learning, Classification, Regression
R&D: Computer-Vision and Video Analytics
S-BigVA Video Analytics
Imagine having to watch a week's worth of motion activated security video of a window with large trees, in the backdrop, that can sometime sway. Task, to find the frames where the intruder is coming through this window. Video Analytics, or video content analysis, is a different problem to motion detection. Unlike motion detection, a scene may be continually changing e.g. something moving in the breeze or background motion yet a fundamental change to the scene (an anomaly) can still be detected. Automatic flagging of anomalous activity within long security cam recordings; including support for multi-camera inference.
Description: Open source, cloud deployable, software for IP cameras with motion-detection. Eliminates false alerts by 70%.
Further information in this white paper.
Status; S-BigVA™Beta version available for DOWNLOAD.
.Software Defined Sensor
SAA software defined sensor S-Sds™ defines an optimal customisation for any standard sensor for specific applications. Applications may range from shock wave detection to low frequency continuous vibrations of electric motors. It also allows use of standard peripherals devices such as line microphones to act as highly sensitive data acquisitions hardware.
Description: S-Sds™ is implemented as software driver and consumer algorithms within S-BigIoT™ and S-BigAnalytics™ respectively.
Applications: Monitoring for predictive-maintenance, shock wave analysis for hydraulic lines, and bio-medical uses.
Remote Object Recognition
Object recognition at very high-magnification with transient atmospheric compensation.
Applications: Hairline crack detection in wind turbines and other relatively inaccessible installations.
Near-orbit object monitoring and detection.
Passive Radio Sensor
Long-range radio beacon tracking analytics.
International Space Station tracking
.
Status; available
Project: High-precision optical system localisation
Objective: Finding precise equipment orientation with IoT motes
Method: Camera enabled reader for fluid level and gyroscopic motion devices embedded in a compact format
Test-bed: Commodity narrow and wide field optical system for field data acquisition coupled with DAC; platform automation with IoT metering and robotic control
Applications:Automatic localisation of surveying systems
Project: Image information retrieval using non-visual analytics
Hypothesis: Some optical sensor data may contain information that is not easily discernible through visual representations.
Objective: Non-visual (informational) artefact discovery from optical sensor data
Method: In-situ data acquisition of naturally graded visual fields with narrow and wide field optical systems; exploration with bio-inspired bigdata analytics
- Noise subtraction
- Associative analyses
- Content-addressable episodic memory
- Contextual data (platform logs, environmental data)
- Data outliers
- CV object-recognition
- Data-object information matrices for ML
Test-bed: Commodity narrow and wide field optical system for field data acquisition coupled with DAC; platform automation with IoT metering and robotic control
Applications: Remote dark object detection, remote site survey, medical diagnostics, autonomous guidance, predictive analysis
* Limited industrial research grants available; EOIs by email below