Isoperla make wildlife products for mobile devices. We make apps for smartphones and tablets, and physical devices built around embedded processors.
Isoperla reinvented wildlife field guides for the mobile age with Automatic Recognition.
This software identifies bird species from songs picked out of the air using the device microphone, like Shazam for birds. The The Times Newspaper said it was an 'astonishingly sophisticated piece of software'. Bird Watching Magazine called it 'remarkably accurate'.
We were the first company in the world to perform Automatic Recognition in real time. We were the first to offer Automatic Recognition on a mobile or consumer device. We are still years ahead of any one else.
We specialise in resource limited devices, mobile platforms such as iOS & Android and Internet of Things (IoT) embedded processors such as Raspberry Pi. We have a comprehensive understanding of the technologies underpinning these systems. We use a blend of agile and traditional techniques to build robust well architected solutions.
We have our own highly functional application framework which supports seamless server integration, data synchronisation, mapping, geolocation and custom user interface controls.
Bird Sond Id for the professional market
The SSR1 is a prototype bird recorder with automatic recognition on board. The device records and stores samples of bird song throughout the day. It analyses the recordings in real-time and lists the species heard. The device can be left outside to record without supervision and collected up to a week later. It is designed for the professional market to help automate bird surveying and will be of interest to ecological consultancies and conservation charities such as the RSPB and BTO.
- The device continuously records 30 second samples of bird song and uses Automatic Recognition to work out which species are singing.
- When an identification is made the result is stored in a spreadsheet on the device and the song sample kept, so results can be validated manually later if required.
- Results can easily be collected from a data stick at the end of the recording period.
- Automatic Recognition uses a standard quantitative multi-variate technique. The application of this method for bird song was developed with Professor of Statistics Jonathan Rougier of Bristol University.
- We do not use classifiers as these have limited accuracy in the natural environment, a multi-variate approach is far more effective. Our method has a much better tolerance of wind, rain, bird proximity and song variability. Classifiers are useful for bats and work well under lab conditions, but are strongly outperformed by a multi-variate approach in the field.
- Recorded samples can be fed back into the Automatic Recognition Learning Machine making it easy to add species.
- At the heart of the device is a Raspberry Pi Zero running the Raspian (Linux) operating system.
- The device does not need a mobile signal or network connection to work, however, it could be made to automatically notify someone if a particular species is detected. This could be via SMS or Email if a suitable mobile signal is available.
- We hope to offer the device for sale from around £250.
We are looking for partners to help take the product to market. Interested ? Drop us an email.
Isoperla was formed in 2007 after receiving a regional innovation award in
association with Bath Spa University and Bath University's Ventures Innovation
Centre. The award was to help develop a new generation of software
for mobile and embedded platforms using multi-variate quantitative statistics.
Wildlife App Range
Isoperla developed the world’s leading brand of wildlife apps, these have sold around $1M worth through the iOS and Android app stores. We sold this brand of apps to our good friends at Sunbird (Mullen and Pohland) in 2016. The films above relate to the products at the point at which we transferred them over to Sunbird. Please check with them about the current status of the apps before purchasing.
|Tree Id||Wild Flower Id||
Bird Song Id