Microalgae Detection by Digital Image Processing and Artificial Intelligence

DOI

Description

This article presents a technical approach to the video computer analysis, to automatically identifying the two most frequently identified microalgae in water supplies. To handle some difficulties encountered in image segmentation problem such as unclear algae boundary and noisy background, we proposed a deep learning-based method for classifiers or localizers to perform microalgae detection and counting process. The system achieves approximately 91% accuracy on Melosira and Oscillatoria detection, which around 4.82 seconds per grid. (Intel Xeon(R) CPU E52667 12 CPU at 2.66GHz and 32.0GB RAM, NVIDIA Quadro K5200 with 2304 CUDA cores). The system can significantly reduce 33.33 - 55.56% of the counting time when compared with the visual inspection of manual methods, and eliminate the error due to the human fatigue.

Journal

Keywords

Details 詳細情報について

  • CRID
    1390578283210981504
  • DOI
    10.5954/icarob.2023.gs2-1
  • ISSN
    21887829
  • Text Lang
    en
  • Data Source
    • JaLC
    • Crossref
  • Abstract License Flag
    Disallowed

Report a problem

Back to top