Building extraction from remote sensing images based on multi-scale information fusion method under Transformer architecture

نویسندگان

چکیده

å»ºç­‘ç‰©æ˜¯åŸŽå¸‚ä¸­æœ€ä¸ºæ™®éçš„åŸºç¡€è®¾æ–½ï¼ŒèŽ·å–é¥æ„Ÿå½±åƒä¸­çš„å»ºç­‘åŒºåŸŸå¯¹äºŽåŸŽå¸‚è§„åˆ’ã€äººå£ä¼°è®¡å’Œç¾æƒ åˆ†æžç­‰å ·æœ‰é‡è¦çš„æ„ä¹‰ã€‚æœ¬æ–‡åŸºäºŽTransformerç»“æž„ï¼Œè®¾è®¡äº†ä¸€ç§ç«¯åˆ°ç«¯çš„é¥æ„Ÿå½±åƒå»ºç­‘åŒºåŸŸæå–æ–¹æ³•ã€‚é¦–å ˆï¼Œé’ˆå¯¹å¤šå°ºåº¦å½±åƒç‰¹å¾å­˜åœ¨çš„ä¿¡æ¯å†—ä½™å’Œä¿¡æ¯å·®å¼‚é—®é¢˜ï¼Œæœ¬æ–‡æå‡ºäº†ä¸€ç§å¤šæ¬¡ç‰¹å¾é‡‘å­—å¡”ç»“æž„Tri-FPN(Triple-Feature Pyramid Networkï¼‰ï¼Œå®žçŽ°è·¨è¶Šè¿‘é‚»å°ºåº¦çš„å ¨å±€å¤šå°ºåº¦ä¿¡æ¯èžåˆï¼Œæé«˜å¤šå°ºåº¦ç‰¹å¾çš„ç±»åˆ«è¡¨å¾ä¸€è‡´æ€§å¹¶å‡å°‘ä¿¡æ¯çš„å†—ä½™ã€‚å ¶æ¬¡ï¼Œé’ˆå¯¹å¤šå°ºåº¦æå–ç»“æžœèžåˆæ—¶ä» è€ƒè™‘å°ºåº¦å› ç´ çš„é—®é¢˜ï¼Œæœ¬æ–‡è®¾è®¡äº†ä¸€ç§é¡¾åŠâ€œå°ºåº¦-类别-空间”的注意力模块CSA-Module(Class-Scale Attention Module),有效融合了不同尺度下的建筑提取结果。最后,在Transformerç»“æž„ä¸ŠåŠ å ¥Tri-FPN与CSA-Module进行模型训练,获得最佳的建筑提取效果。实验对比分析表明,本文的方法有效提高了建筑区域的检出率,并提供出更为准确的建筑轮廓,提升了遥感影像中建筑的提取精度,在WHU Building数据集和INRIA数据集上分别取得了91.53%和81.7%的IOU分数。

برای دانلود باید عضویت طلایی داشته باشید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A detection method of artificial area from high resolution remote sensing images based on multi scale and multi feature fusion

In order to solve the problem of automatic detection of artificial objects in high resolution remote sensing images, a method for detection of artificial areas in high resolution remote sensing images based on multi-scale and multi feature fusion is proposed. Firstly, the geometric features such as corner, straight line and right angle are extracted from the original resolution, and the pseudo ...

متن کامل

Object-oriented Change Detection for Remote Sensing Images Based on Multi-scale Fusion

In the process of object-oriented change detection, the determination of the optimal segmentation scale is directly related to the subsequent change information extraction and analysis. Aiming at this problem, this paper presents a novel object-level change detection method based on multi-scale segmentation and fusion. First of all, the fine to coarse segmentation is used to obtain initial obje...

متن کامل

Road Extraction from High Resolution Remote Sensing Images based on Multi-features and Multi-stages

Road extraction from high resolution remote sensing imagery is very important for many applications such as GIS data updating, transportation management and city planning. In this paper we propose a semi-automatic multi-stage method to extract roads from high resolution remote sensing imagery based on multi-features. The proposed method contains two main steps, i.e. segmenting original image wi...

متن کامل

Extraction Method for Earthquake-Collapsed Building Information Based on High-Resolution Remote Sensing

At present, the extraction of earthquake disaster information from remote sensing data relies on visual interpretation. However, this technique cannot effectively and quickly obtain precise and efficient information for earthquake relief and emergency management. Collapsed buildings in the town of Zipingpu after the Wenchuan earthquake were used as a case study to validate two kinds of rapid ex...

متن کامل

Information Mining from Remote Sensing Imagery Based on Multi-scale and Multi-feature Processing Techniques

The paper attempts to present an information extraction approach in terms of image segmentation based on an object-oriented algorithm for high-resolution remote sensing images. The hierarchy frame and multi-features of the remote sensing image understanding and processing method are put forward. Firstly we extract various internal features of relatively homogeneous primitive objects using an im...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

ژورنال

عنوان ژورنال: Journal of remote sensing

سال: 2023

ISSN: ['1007-4619', '2095-9494']

DOI: https://doi.org/10.11834/jrs.20233017