DEVELOPMENT OF A HABITAT PREDICTION METHOD FOR RUDITAPES PHILIPPINARUM BASED ON IMAGE ANALYSIS USING DEEP LEARNING

DOI Web Site 1 Citations 2 References

Bibliographic Information

Other Title
  • UAV空撮画像を用いた機械学習によるアサリの生息場予測手法の開発

Abstract

<p> The present study predicted the occurrence of the Japanese Littleneck Ruditapes philippinarum using presence/absence data from filed surveys and image analysis of mud flat surface employing deep learning around the river mouth of the Fushino River in the Yamaguchi Bay, Yamaguchi Prefecture, where the decline in the catch of the species is an issue. The shells were found inhabiting at 28 out of the 193 surveyed sites in the Yamaguchi Bay. Of the 448 pictures, each from present or absent sites, 69.5% were judged correctly to be present or absent in total by deep learning. Although pictures taken at sites where the species is apparently absent (i.e. on top of roads, sandy beaches, or rocks) were judged correctly, the algorithm tended to misjudge absent sites as present. The results indicate that even though the program is currently not practicable, analyzing pictures taken by UAV with deep learning could be usable to grasp the general trends in the habitat of the species.</p>

Journal

Citations (1)*help

See more

References(2)*help

See more

Keywords

Details 詳細情報について

Report a problem

Back to top