β

Automating Wildlife Image Processing Using IoT and

Intel® Developer Zone 11 阅读
 
English
Automating Wildlife Image Processing Using IoT and the Intel® Movidius™ Neural Compute Stick

Where's The Bear (WTB) implements a multi-tier (cloud, edge, sensing) system that integrates recent advances in machine learning based image processing to automatically classify animals in images from remote, motion-triggered camera traps. It uses non-local, resource-rich, public/private cloud systems to train the machine learning models, and "in-the-field,'' resource-constrained edge systems to perform classification near the IoT sensing devices (cameras). WTB is deployed at the UCSB Sedgwick Reserve, a 6000 acre site for environmental research and used it to aggregate, manage, and analyze over 1.12 M images. WTB integrates TensorFlow* and OpenCV applications to perform automatic classification and tagging for a subset of these images. To avoid transferring large numbers of training images for TensorFlow over a low-bandwidth network linking Sedgwick to the public/private clouds, a technique was devised that uses stock Google Images to construct a synthetic training set using only a small number of empty, background images from Sedgwick. The system is able to accurately identify bears, deer, coyotes, and empty images and significantly reduces the time and bandwidth requirements for image transfer, as well as end-user analysis time, since WTB automatically filters the images on-site.

  • Intel® Student Ambassadors
  • Students
  • Artificial Intelligence
  • Include in RSS:

    1

    Overview:

    Where's The Bear (WTB) implements a multi-tier (cloud, edge, sensing) system that integrates recent advances in machine learning based image processing to automatically classify animals in images from remote, motion-triggered camera traps. It uses non-local, resource-rich, public/private cloud systems to train the machine learning models, and "in-the-field,'' resource-constrained edge systems to perform classification near the IoT sensing devices (cameras). WTB is deployed at the UCSB Sedgwick Reserve, a 6000 acre site for environmental research and used it to aggregate, manage, and analyze over 1.12 M images. WTB integrates TensorFlow* and OpenCV applications to perform automatic classification and tagging for a subset of these images. To avoid transferring large numbers of training images for TensorFlow over a low-bandwidth network linking Sedgwick to the public/private clouds, a technique was devised that uses stock Google Images to construct a synthetic training set using only a small number of empty, background images from Sedgwick. The system is able to accurately identify bears, deer, coyotes, and empty images and significantly reduces the time and bandwidth requirements for image transfer, as well as end-user analysis time, since WTB automatically filters the images on-site.

    Video Duration (in milliseconds):

    83400

    Subtitle Background Opacity:

    100.00

    Default to English Subtitles?:

     
    作者:Intel® Developer Zone
    原文地址:Automating Wildlife Image Processing Using IoT and, 感谢原作者分享。

    发表评论