Advancing Technology
&
Gaining Efficiency
Work Smarter Not Harder
Technology for Cost-Effective Fisheries Sampling
In 2014, MER began developing multiple technologies to help with the implementation of large scale sampling projects in the US Caribbean. The development project started with a simple electronic reporting platform which grew to include image capture and a module to link automated weight recording. Technology has advanced to the point where big data no longer costs a fortune to handle and very advanced machine learning and artificial intelligence allows for automated species ID and for automated length measurements. Our sampling stations document every fish via an image or video and allows for verification and high quality data. Only minimal training is required making fisheries data collection accessible to individuals from diverse cultural or educational backgrounds.
Reporting, oversight, quality control, and image capture for ID verification and from Smart Scale
Simple at-sea video capture system
Automated image and weight capture
Species identification and length verification as part of algorithm tuning process
MER has partnered with the team that developed MERLIN (bird image and call/audio identification application) at The Cornell Lab of Ornithology to handle this big data challenge.
E-Data Android Platform
Our E-Data Platform is a simple, intuitive, easy to use android application. We deployed on $150 Samsung tablets that had decent cameras. Data are stored locally on the tablet, quality controlled on entry, and uploaded to database when sampler has access to wi-fi.
Supervisors had access to GPS locations of all tablets allowing for oversight and verification of assignment completion.
Smart Scale - Prototype 1
AI Assisted Port Sampling Stations
"Smart Scale"
Upon stabilization the system, which is directly integrated into the e-data platform, automatically takes a picture and records the weight of each individual. Automated length measurements are calculated upon upload via artificial intelligence (in development).
Electronic Observers
Remote Electronic monitoring
MER developed an at-sea video capture system which has been deployed in Puerto Rico to capture the results of a lobster selectivity Study. With the push of a single button, fishers start both a deck camera and a image capture camera. Each expermental treatement is indicated by a different colour tag. Post processing to automate object extraction and carapace measurements are in development.
Our Annotation software is being developed using Python and a basic spectral analysis of each photo to suggest species. Annotation points are indicated by red circles and suggested species are in the right thumbnails.
Annotation Software
Machine Learning/
Artificial Intelligence -
A.I. Assisted Sampling
