<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.4 20241031//EN"
        "https://jats.nlm.nih.gov/publishing/1.4/JATS-journalpublishing1-4.dtd">
<article  article-type="research-article"        dtd-version="1.4">
            <front>

                <journal-meta>
                                                                <journal-id>demiryolu mühendisliği</journal-id>
            <journal-title-group>
                                                                                    <journal-title>Demiryolu Mühendisliği</journal-title>
            </journal-title-group>
                            <issn pub-type="ppub">2149-1607</issn>
                                        <issn pub-type="epub">2687-2463</issn>
                                                                                            <publisher>
                    <publisher-name>Demiryolu Mühendisleri Derneği</publisher-name>
                </publisher>
                    </journal-meta>
                <article-meta>
                                        <article-id pub-id-type="doi">10.47072/demiryolu.1336812</article-id>
                                                                <article-categories>
                                            <subj-group  xml:lang="en">
                                                            <subject>Data Communications</subject>
                                                    </subj-group>
                                            <subj-group  xml:lang="tr">
                                                            <subject>Veri İletişimleri</subject>
                                                    </subj-group>
                                    </article-categories>
                                                                                                                                                        <title-group>
                                                                                                                        <trans-title-group xml:lang="en">
                                    <trans-title>Comparative Analysis of Deep Learning-Based Methods for Making Sense of Railway and Its Environment</trans-title>
                                </trans-title-group>
                                                                                                                                                                                                <article-title>Demiryolu Ray ve Çevresinin Anlamlandırılması için Derin Öğrenme Tabanlı Yöntemlerin Karşılaştırmalı Analizi</article-title>
                                                                                                    </title-group>
            
                                                    <contrib-group content-type="authors">
                                                                        <contrib contrib-type="author">
                                                                    <contrib-id contrib-id-type="orcid">
                                        https://orcid.org/0000-0001-6880-4935</contrib-id>
                                                                <name>
                                    <surname>Aydın</surname>
                                    <given-names>İlhan</given-names>
                                </name>
                                                                    <aff>FIRAT ÜNİVERSİTESİ</aff>
                                                            </contrib>
                                                    <contrib contrib-type="author">
                                                                    <contrib-id contrib-id-type="orcid">
                                        https://orcid.org/0000-0002-9846-967X</contrib-id>
                                                                <name>
                                    <surname>Şener</surname>
                                    <given-names>Taha Kubilay</given-names>
                                </name>
                                                                    <aff>FIRAT UNIVERSITY</aff>
                                                            </contrib>
                                                    <contrib contrib-type="author">
                                                                    <contrib-id contrib-id-type="orcid">
                                        https://orcid.org/0000-0001-6952-8880</contrib-id>
                                                                <name>
                                    <surname>Sevi</surname>
                                    <given-names>Mehmet</given-names>
                                </name>
                                                                    <aff>MUŞ ALPARSLAN ÜNİVERSİTESİ</aff>
                                                            </contrib>
                                                                                </contrib-group>
                        
                                        <pub-date pub-type="pub" iso-8601-date="20240131">
                    <day>01</day>
                    <month>31</month>
                    <year>2024</year>
                </pub-date>
                                                    <issue>19</issue>
                                        <fpage>1</fpage>
                                        <lpage>16</lpage>
                        
                        <history>
                                    <date date-type="received" iso-8601-date="20230802">
                        <day>08</day>
                        <month>02</month>
                        <year>2023</year>
                    </date>
                                                    <date date-type="accepted" iso-8601-date="20230914">
                        <day>09</day>
                        <month>14</month>
                        <year>2023</year>
                    </date>
                            </history>
                                        <permissions>
                    <copyright-statement>Copyright © 2014, Demiryolu Mühendisliği</copyright-statement>
                    <copyright-year>2014</copyright-year>
                    <copyright-holder>Demiryolu Mühendisliği</copyright-holder>
                </permissions>
            
                                                                                                <trans-abstract xml:lang="en">
                            <p>Rail safety in railways is very important for the prevention of train accidents. Objects around and on the rails pose a danger to the train. Therefore, the safe operation of trains by detecting unauthorized access to the railway is very important for smart transportation systems. In this study, image segmentation-based approaches are compared in order to make sense of the rail environment in railway systems, and the objects around the rail are detected. UNet, BiseNetV2, DeepLabV3, and PP-LiteSeg models were analyzed comparatively to segment of the rail and its environment based on image segmentation. In addition, YOLOv7 has been applied to detect objects around the rail. Thus, it was evaluated how successful the models were in real-world scenarios. As a result of the experiments, it was determined that the PP-LiteSeg model, which stands out with its lightweight structure, showed high segmentation performance. It has been seen that the training phase is important in object detection, and it has been concluded that PP-LiteSeg can be successfully applied on single circuit boards such as Jetson Nano. Another model in the study, YOLOv7, has been optimized to run in parallel using the TensorRT library. A special control mechanism has been developed to use memory areas independently. According to the results obtained, it was seen that the PP-LiteSeg model achieved higher accuracy and mIoU values than other models. The study includes important results for the selection of segmentation models for fast and accurate object detection in rail systems. The study proved that with the use of the PP-LiteSeg model, high-quality object detection can be achieved even in environments with limited resources.</p></trans-abstract>
                                                                                                                                    <abstract><p>Demiryollarında ray güvenliği tren kazalarının önlenmesi için oldukça önemlidir. Ray çevresinde ve üzerinde bulunan nesneler tren için tehlike arz etmektedir. Dolayısıyla demiryoluna izinsiz girişlerin tespit edilerek trenlerin güvenli çalışması akıllı ulaşım sistemleri için oldukça önemlidir. Bu çalışmada raylı sistemlerde ray çevresinin anlamlandırılması amacıyla görüntü bölütleme tabanlı yaklaşımlar karşılaştırılmış ve ray çevresindeki nesnelerin tespiti sağlanmıştır. Görüntü bölütleme tabanlı ray ve çevresinin anlamlandırılması için UNet, BiSeNetV2, DeepLabV3 ve PP-LiteSeg modelleri karşılaştırmalı olarak analiz edilmiştir. Ayrıca ray çevresindeki nesnelerin tespitinde YOLOv7 uygulanmıştır. Böylece, modellerin gerçek dünya senaryolarında ne kadar başarılı olduğu değerlendirilmiştir. Deneyler sonucunda, hafif yapısıyla dikkat çeken PP-LiteSeg modelinin yüksek segmentasyon performansı gösterdiği tespit edilmiştir. Eğitim aşamasının nesne tespitinde önemli olduğu görülmüş ve PP-LiteSeg&#039;in Jetson Nano gibi tek devre kartlarda başarılı bir şekilde uygulanabildiği sonucuna ulaşılmıştır. Çalışmadaki bir diğer model YOLOv7, TensorRT kütüphanesi kullanılarak paralel çalışacak şekilde optimize edilmiş ve hafıza alanlarının bağımsız olarak kullanılabilmesi için özel bir kontrol mekanizması geliştirilmiştir. Elde edilen sonuçlara göre, PP-LiteSeg modelinin diğer modellere göre daha yüksek doğruluk ve mIoU değerleri elde ettiği görülmüştür. Yapılan çalışma raylı sistemlerde hızlı ve doğru nesne tespiti için segmentasyon modellerinin seçimine yönelik önemli sonuçlar içermektedir. Çalışma PP-LiteSeg modelinin kullanımıyla birlikte sınırlı kaynağa sahip ortamlarda bile yüksek kalitede nesne tespiti yapılabileceğini kanıtlamıştır.</p></abstract>
                                                            
            
                                                                                        <kwd-group>
                                                    <kwd>Derin öğrenme</kwd>
                                                    <kwd>  Demiryolu</kwd>
                                                    <kwd>  Nesne tespiti</kwd>
                                                    <kwd>  Akıllı ulaşım</kwd>
                                                    <kwd>  YOLO</kwd>
                                                    <kwd>  Semantik segmentasyon</kwd>
                                            </kwd-group>
                            
                                                <kwd-group xml:lang="en">
                                                    <kwd>Deep learning</kwd>
                                                    <kwd>  Railway</kwd>
                                                    <kwd>  Object detection</kwd>
                                                    <kwd>  Intelligent transportation</kwd>
                                                    <kwd>  YOLO</kwd>
                                                    <kwd>  Semantic segmentation</kwd>
                                            </kwd-group>
                                                                                                                                    <funding-group specific-use="FundRef">
                    <award-group>
                                                    <funding-source>
                                <named-content content-type="funder_name">Fırat Üniversitesi Bilimsel Araştırma Projeleri Birimi</named-content>
                            </funding-source>
                                                                            <award-id>ADEP.22.02</award-id>
                                            </award-group>
                </funding-group>
                                </article-meta>
    </front>
    <back>
                            <ref-list>
                                    <ref id="ref1">
                        <label>1</label>
                        <mixed-citation publication-type="journal">[1]	T. Zhu, &amp; JMMS. De Pedro, ”Railway traffic conflict detection via a state transition prediction approach,” IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 5, pp. 1268-1278.</mixed-citation>
                    </ref>
                                    <ref id="ref2">
                        <label>2</label>
                        <mixed-citation publication-type="journal">[2]	T. Ye, Z. Zheng, X. Li, Z. Zhao, &amp; XZ. Gao, “An efficient few-shot object detection method for railway intrusion via fine-tune approach and contrastive learning,” IEEE Transactions on Instrumentation and Measurement.</mixed-citation>
                    </ref>
                                    <ref id="ref3">
                        <label>3</label>
                        <mixed-citation publication-type="journal">[3]	Y. Li, Y. Qin, Z. Xie, Z. Cao, L. Jia, Z. Yu &amp; E. Zhang, “Efficient SSD: a real-time intrusion object detection algorithm for railway surveillance,” In 2020 International Conference on Sensing, Diagnostics, Prognostics, and Control (SDPC), pp. 391-395. IEEE</mixed-citation>
                    </ref>
                                    <ref id="ref4">
                        <label>4</label>
                        <mixed-citation publication-type="journal">[4]	Z. Zheng, W. Liu, R. Liu, L. Wang, L. Mao, Q. Qiu, &amp; G. Ling,  “Anomaly detection of metro station tracks based on sequential updatable anomaly detection framework,” IEEE Transactions on Circuits and Systems for Video Technology,  vol. 32, no. 11, pp. 7677-7691</mixed-citation>
                    </ref>
                                    <ref id="ref5">
                        <label>5</label>
                        <mixed-citation publication-type="journal">[5]	X. Ding, X. Cai, Z. Zhang, W. Liu, &amp; W. Song, “Railway foreign object intrusion detection based on deep learning,” In 2022 International Conference on Computer Engineering and Artificial Intelligence (ICCEAI), pp. 735-739. IEEE.</mixed-citation>
                    </ref>
                                    <ref id="ref6">
                        <label>6</label>
                        <mixed-citation publication-type="journal">[6]	SS. Kırat &amp; İ. Aydın, &quot;Açıklanabilir Yapay Zekâ Tabanlı Denetimsiz Öğrenme ile Ray Kusur Tespiti&quot;, Demiryolu Mühendisliği, vol. 18, pp. 1-13. doi:10.47072/demiryolu.1231751.</mixed-citation>
                    </ref>
                                    <ref id="ref7">
                        <label>7</label>
                        <mixed-citation publication-type="journal">[7]	M. Sevi, İ. Aydın, E. Akın, “Detection of rail surface defects based on ensemble learning of YOLOv5,” Demiryolu Mühendisliği, vol. 17, pp. 115-132. doi: 10.47072/demiryolu.1205483.</mixed-citation>
                    </ref>
                                    <ref id="ref8">
                        <label>8</label>
                        <mixed-citation publication-type="journal">[8]	D. Çetintaş &amp; T. Tuncer, “Determining the type of document read using eye movement properties by hybrid CNN method,” Traitement du Signal, vol. 39, no. 4, pp. 1099.</mixed-citation>
                    </ref>
                                    <ref id="ref9">
                        <label>9</label>
                        <mixed-citation publication-type="journal">[9]	X. Gong, X, Chen, Z. Zhong &amp; W. Chen, “Enhanced few-shot learning for intrusion detection in railway video surveillance,” IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 8, pp. 11301-11313.</mixed-citation>
                    </ref>
                                    <ref id="ref10">
                        <label>10</label>
                        <mixed-citation publication-type="journal">[10]	X. Li, L. Zhu, Z. Yu, B. Guo, &amp; Y. Wan, “Vanishing point detection and rail segmentation based on deep multi-task learning,” IEEE Access, vol. 8, pp. 163015-163025.</mixed-citation>
                    </ref>
                                    <ref id="ref11">
                        <label>11</label>
                        <mixed-citation publication-type="journal">[11]	X. Ding, X. Cai, Z. Zhang, W. Liu, &amp; W. Song, “Railway foreign object intrusion detection based on deep learning,” In 2022 International Conference on Computer Engineering and Artificial Intelligence (ICCEAI), pp. 735-739. IEEE.</mixed-citation>
                    </ref>
                                    <ref id="ref12">
                        <label>12</label>
                        <mixed-citation publication-type="journal">[12]	H. Huang, G. Zhao, Y. Bo, J. Yu, L. Liang, Y. Yang, &amp; K. Ou, ”Railway intrusion detection based on refined spatial and temporal features for UAV surveillance scene,” Measurement, vol. 211, 112602.</mixed-citation>
                    </ref>
                                    <ref id="ref13">
                        <label>13</label>
                        <mixed-citation publication-type="journal">[13]	Z. Cao, Y. Qin, Z.  Xie, Q. Liu, E. Zhang, Z. Wu &amp; Z. Yu, ”An effective railway intrusion detection method using dynamic intrusion region and lightweight neural network,” Measurement, vol. 191, 110564.</mixed-citation>
                    </ref>
                                    <ref id="ref14">
                        <label>14</label>
                        <mixed-citation publication-type="journal">[14]	M. Sevi, İ. Aydın, “Improving Unet segmentation performance using an ensemble model in images containing railway lines,” Turk J Elec Eng &amp; Comp Sci,  vol. 34, no. 4, pp. 739-750.</mixed-citation>
                    </ref>
                                    <ref id="ref15">
                        <label>15</label>
                        <mixed-citation publication-type="journal">[15]	O. Ronneberger, P. Fischer, &amp; T. Brox, “U-net: convolutional networks for biomedical image segmentation,” In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, 2015, Proceedings, Part III 18, pp. 234-241.</mixed-citation>
                    </ref>
                                    <ref id="ref16">
                        <label>16</label>
                        <mixed-citation publication-type="journal">[16]	LC. Chen, G. Papandreou, F. Schroff &amp; H. Adam, “Rethinking atrous convolution for semantic image segmentation,” arXiv preprint arXiv:1706.05587.</mixed-citation>
                    </ref>
                                    <ref id="ref17">
                        <label>17</label>
                        <mixed-citation publication-type="journal">[17]	C. Yu, C. Gao, J. Wang, G. Yu, C. Shen &amp; N. Sang, “Bisenet v2: Bilateral network with guided aggregation for real-time semantic segmentation,” International Journal of Computer Vision, vol. 129, pp. 3051-3068.</mixed-citation>
                    </ref>
                                    <ref id="ref18">
                        <label>18</label>
                        <mixed-citation publication-type="journal">[18]	J. Peng, Y. Liu, S. Tang, Y.  Hao, L. Chu, G. Chen, &amp; Y. Ma, “Pp-liteseg: A superior real-time semantic segmentation model,” arXiv preprint arXiv:2204.02681.</mixed-citation>
                    </ref>
                                    <ref id="ref19">
                        <label>19</label>
                        <mixed-citation publication-type="journal">[19]	CY. Wang, A. Bochkovskiy &amp; HYM. Liao, “YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,” arXiv 2022. arXiv preprint arXiv:2207.02696.</mixed-citation>
                    </ref>
                                    <ref id="ref20">
                        <label>20</label>
                        <mixed-citation publication-type="journal">[20]	O. Zendel, M. Murschitz, M. Zeilinger, D. Steininger, S. Abbasi, S &amp; C. Beleznai, “Railsem19: a dataset for semantic rail scene understanding,” In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pp. 0-0.</mixed-citation>
                    </ref>
                                    <ref id="ref21">
                        <label>21</label>
                        <mixed-citation publication-type="journal">[21]	J. Redmon, S. Divvala, R. Girshick &amp; A. Farhadi, “You only look once: Unified, real-time object detection,” In Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 779-788.</mixed-citation>
                    </ref>
                                    <ref id="ref22">
                        <label>22</label>
                        <mixed-citation publication-type="journal">[22]	P. Li, H. Xiong, J. Fan, “Sun dfanet: Deep feature aggregation for real-time semantic segmentation,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2019, pp. 9522-9531</mixed-citation>
                    </ref>
                                    <ref id="ref23">
                        <label>23</label>
                        <mixed-citation publication-type="journal">[23]	E. Romera, JM. Alvarez, LM. Bergasa, R. Arroyo, “Erfnet: efficient residual factorized convnet for real-time semantic segmentation,” IEEE Transactions on Intelligent Transportation Systems 2017: 19, pp. 263-272.</mixed-citation>
                    </ref>
                                    <ref id="ref24">
                        <label>24</label>
                        <mixed-citation publication-type="journal">[24]	RP. Poudel, S. Liwicki, R. Cipolla, “Fast-scnn: fast semantic segmentation network,” arXiv preprint arXiv:1902.04502 2019.</mixed-citation>
                    </ref>
                                    <ref id="ref25">
                        <label>25</label>
                        <mixed-citation publication-type="journal">[25]	H. Zhao, X. Qi, X. Shen, J. Shi &amp; J. Jia, “Icnet for real-time semantic segmentation on high-resolution images,” In: Proc. European Conference on Computer Vision (ECCV) 2018, pp. 405– 420.</mixed-citation>
                    </ref>
                            </ref-list>
                    </back>
    </article>
