RFP: Machine Vision for Speciation and Enumeration of Discarded Catch in the West Coast Groundfish Trawl Fishery

EM Info Admin

Published April 29, 2019
  • EDF is seeking a suitably qualified technology and services provider to work with us and our fishermen partners to design, implement, and test a ‘smart’ catch handling protocol in the West Coast groundfish bottom trawl fishery that focuses on the development of a machine vision algorithm that is capable of identifying and counting discarded species in the fishery. Working with one fisherman partner based out of Warrenton, Oregon, the provider will co-design a catch handling protocol which will enable the integration of a camera-based machine vision system into the process. The project partner already has a working conveyor belt that is part of the existing catch handling process and we therefore expect there to be minimal alteration of this physical process. However, any modifications that need to be made to the speed of the belt, the lighting conditions, the depth and orientation of species placed on the belt, and any other aspects of the process, we expect will be the main subject of the co-design process. The camera system will be installed by the provider but will not be integrated with the EM system on the vessel. All data transmission (via hard drive transfer or wirelessly), storage, and analysis (labeling for the purposes of training the machine vision algorithm) of video data will be the responsibility of the provider.

    The provider will then engage in the development of a machine vision algorithm that is capable of automatically identifying and counting discarded species that go over the conveyor belt to a high level of accuracy. A quantitative test of the performance of this algorithm will be conducted over an extended period of time and all intermediate and final results will be made available to project partners. In addition, we expect that the algorithm developed during the course of this project will be made freely and publicly available to help future development in this field.

    Download full RFP

 

Leave a Reply

 

Your email address will not be published. Required fields are marked *