Automated Identification and Counting of Saigas (Saiga tatarica) by Using Deep Convolutional Neural Networks in High-Resolution Satellite Images

IF 0.5 4区生物学 Q4 BIOLOGY

Biology Bulletin Pub Date : 2024-09-10 DOI:10.1134/s1062359024608784

V. V. Rozhnov, A. L. Salman, A. A. Yachmennikova, A. A. Lushchekina, P. A. Salman

{"title":"Automated Identification and Counting of Saigas (Saiga tatarica) by Using Deep Convolutional Neural Networks in High-Resolution Satellite Images","authors":"V. V. Rozhnov, A. L. Salman, A. A. Yachmennikova, A. A. Lushchekina, P. A. Salman","doi":"10.1134/s1062359024608784","DOIUrl":null,"url":null,"abstract":"<h3 data-test=\"abstract-sub-heading\">Abstract</h3><p>We utilized a two-phase analysis using deep convolutional neural networks (DCNN) to create an automated technology that enabled us to detect and count saigas (<i>Saiga tatarica</i>) in satellite images with a resolution of 0.3–0.5 m/pixel (Eros-B 2012; 2013 and Beijing KA 2022 satellites). In the first phase, the satellite image is automatically divided into sections and checked for the presence or absence of clusters of objects (the “classification” phase). Then, during the second phase, only the fragments of the satellite image where at least one saiga was previously found are analyzed (the “detection” phase). The method was calibrated by training a neural network on the results of the preliminary processing of archival satellite images from 2012 and 2013, carried out manually by zoological experts. When we tested the DCNN work with a “confidence threshold” of 0.3, we identified 1284 saigas on the entire model satellite image, while a zoological expert manually identified 1412 saigas. For practical use and to assess the effectiveness of this method, we counted saigas on a 2022 image covering two adjacent specially protected natural areas (PAs) located in the Republic of Kalmykia and the Astrakhan region (Russian Federation). The results are presented with different “thresholds of confidence.”</p>","PeriodicalId":55366,"journal":{"name":"Biology Bulletin","volume":"28 1","pages":""},"PeriodicalIF":0.5000,"publicationDate":"2024-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Biology Bulletin","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.1134/s1062359024608784","RegionNum":4,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"BIOLOGY","Score":null,"Total":0}

引用次数: 0

Abstract

We utilized a two-phase analysis using deep convolutional neural networks (DCNN) to create an automated technology that enabled us to detect and count saigas (Saiga tatarica) in satellite images with a resolution of 0.3–0.5 m/pixel (Eros-B 2012; 2013 and Beijing KA 2022 satellites). In the first phase, the satellite image is automatically divided into sections and checked for the presence or absence of clusters of objects (the “classification” phase). Then, during the second phase, only the fragments of the satellite image where at least one saiga was previously found are analyzed (the “detection” phase). The method was calibrated by training a neural network on the results of the preliminary processing of archival satellite images from 2012 and 2013, carried out manually by zoological experts. When we tested the DCNN work with a “confidence threshold” of 0.3, we identified 1284 saigas on the entire model satellite image, while a zoological expert manually identified 1412 saigas. For practical use and to assess the effectiveness of this method, we counted saigas on a 2022 image covering two adjacent specially protected natural areas (PAs) located in the Republic of Kalmykia and the Astrakhan region (Russian Federation). The results are presented with different “thresholds of confidence.”

Abstract Image

查看原文本刊更多论文

在高分辨率卫星图像中使用深度卷积神经网络自动识别和计算 "斋加"（Saiga tatarica）数量

摘要我们利用深度卷积神经网络（DCNN）进行了两阶段分析，创建了一种自动化技术，使我们能够在分辨率为 0.3-0.5 m/pixel 的卫星图像（Eros-B 2012；2013 和 Beijing KA 2022 卫星）中检测和计数 "箭猪"（Saiga tatarica）。在第一阶段，卫星图像被自动分割成若干部分，并检查是否存在物体群（"分类 "阶段）。然后，在第二阶段，只分析之前至少发现过一个赛加的卫星图像片段（"检测 "阶段）。动物学专家对 2012 年和 2013 年的档案卫星图像进行了初步处理，并根据处理结果对神经网络进行了训练，从而对该方法进行了校准。当我们以 0.3 的 "置信度阈值 "对 DCNN 的工作进行测试时，我们在整个模型卫星图像上识别出了 1284 个红鼠，而动物学专家手动识别出了 1412 个红鼠。为了实际使用和评估该方法的有效性，我们在一幅 2022 年的图像上对红鼠进行了计数，该图像覆盖了位于卡尔梅克共和国和阿斯特拉罕州（俄罗斯联邦）的两个相邻的特别保护自然区（PA）。结果以不同的 "置信阈值 "呈现。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Biology Bulletin 生物-生物学

CiteScore

0.70

自引率

20.00%

发文量

审稿时长

4-8 weeks

期刊介绍： Biology Bulletin (Izvestiya Rossiiskoi Akademii Nauk – Seriya Biologicheskaya) is an interdisciplinary journal of general biology. It focuses on fundamental studies in the fields of cell biology, biochemistry, zoology, botany, physiology, and ecology. This journal publishes current materials of experimental studies and surveys on current problems in general biology. It also publishes information on scientific conferences and new books in the fields of general biology.