AI画像解析による汽水性希少ハゼ類の生息状況の予測手法の開発
書誌事項
- タイトル別名
-
- DEVELOPMENT OF PREDICTION METHOD FOR HABITAT OF BRACKISHWATER RARE GOBIES USING AI IMAGE ANALYSIS
抄録
<p>It is necessary to develop an efficient habitat survey method that does not depend on the surveyor's ability to collect organisms in the tidal flats. We have developed a method to predict the habitat status of goby fishes inhabiting tidal flats by using image analysis, which is one of the specialties of convolutional neural networks, a type of deep learning. The percentage of correct answers for the three species of goby species was relatively high, ranging from 0.72 to 0.74 and the accuracy of identifying them as “absent” was extremely high, ranging from 0.81 to 0.94. This result indicates that AI image analysis makes a significant contribution to the efficiency of habitat surveys in tidal flats.</p>
収録刊行物
-
- 河川技術論文集
-
河川技術論文集 26 (0), 373-378, 2020
公益社団法人 土木学会
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390854717782249088
-
- ISSN
- 24366714
-
- 本文言語コード
- ja
-
- データソース種別
-
- JaLC
-
- 抄録ライセンスフラグ
- 使用可