{"title":"Object detection of linear structures on Mars based on YOLO network","authors":"Jingkun Xu, Jiarui Liang, Pengcheng Yan, Xiaolin Tian","doi":"10.1145/3579654.3579750","DOIUrl":null,"url":null,"abstract":"The geographic feature of terrestrial planets is a critical and important reference that could help researchers have a further understanding of planetary history and its evo-lution. Traditionally, the detection of specific landforms and their geographic parameter extraction basically relies on manual marking, these types of approach may consume a lot of labor and time costs. On the other side, with the development of convolutional neural networks (CNNs),it is able to handle more complicated tasks such as object detection and semantic segmentation with high efficiency and accuracy. In order to solve this problem, this paper presents an introduction about using neural network to do object detection of the linear structures on Mars, like dorsum, fossa and so on. Based on the DEM data of Mars, this paper makes a linear structure data set. The neural network be used is YOLO-v5. In the test of 300 iterations, the algorithm can get the best detection results when it iterates 200 times, the accuracy of the object detection can reach 81% when mAP = 0.5. The results show that the method proposed in this paper can effectively judge whether there is a linear structure in the graph and mark it, which can be used to assist scientists to reduce the time cost required for detection.","PeriodicalId":146783,"journal":{"name":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","volume":"60 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 2022 5th International Conference on Algorithms, Computing and Artificial Intelligence","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3579654.3579750","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
The geographic feature of terrestrial planets is a critical and important reference that could help researchers have a further understanding of planetary history and its evo-lution. Traditionally, the detection of specific landforms and their geographic parameter extraction basically relies on manual marking, these types of approach may consume a lot of labor and time costs. On the other side, with the development of convolutional neural networks (CNNs),it is able to handle more complicated tasks such as object detection and semantic segmentation with high efficiency and accuracy. In order to solve this problem, this paper presents an introduction about using neural network to do object detection of the linear structures on Mars, like dorsum, fossa and so on. Based on the DEM data of Mars, this paper makes a linear structure data set. The neural network be used is YOLO-v5. In the test of 300 iterations, the algorithm can get the best detection results when it iterates 200 times, the accuracy of the object detection can reach 81% when mAP = 0.5. The results show that the method proposed in this paper can effectively judge whether there is a linear structure in the graph and mark it, which can be used to assist scientists to reduce the time cost required for detection.