Generalization ability of cloud detection network for satellite imagery based on DeepLabv3+
نویسندگان
چکیده
è¿?å¹´æ¥ï¼?深度å¦ä¹ ç®?æ³?å¾?å?°äº?é?¿è¶³ç??å?å±?ï¼?并å¼?å§?åº?ç?¨äº?äº?æ£?æµ?ã??ä½?æ?¯æ·±åº¦ç¥?ç»ç½?ç»?模å??å?æ?°ä¼?å¤?ï¼?ä¾èµ?大é?è®ç»?æ ·æ?¬ï¼?å? æ¤ç?è§£å ¶æ³?å??æ?§è?½å¯¹äº?深度å¦ä¹ å?¨ä¸å?é¥æ??å½±å?ç??å®?é? åº?ç?¨å ·æ??é?è¦ç??å?è??ä»·å?¼ã??æ?¬æ??以深度è¯ä¹?å??å?²ç®?æ³?DeepLabv3+为ä¾?ï¼?以ä¸?ç»?广æ³?使ç?¨ç??äº?æ ?è®°æ?°æ®é??â??L8 Biomeâ?为éª?è¯æ?°æ®ï¼?æ?¢è®¨è¯¥ç®?æ³?ç?¨äº?äº?æ£?æµ?æ?¶å?¨ä¸å?ç??å?°è¡¨æ?¯è§?ã?空é?´å??辨ç??å??å ?谱波段ç»?å?ç??é¥æ??å½±å?ä¸?ç??æ³?å??æ?§è?½ã??äº?æ Biomeâ?å? å«96æ?¯å ·æ??å ¨ç?代表æ?§ç??Landsat 8 OLIå½±å?å?ç?¸åº?ç??人工äº?æ?©è??ï¼?被广æ³?ç?¨äº?æµ?è¯?äº?å??äº?é?´å½±æ£?æµ?ç®?æ³?æ?§è?½ã??é¦?å ?ï¼?å?©ç?¨Landsat OLIäº?æ Biomeâ?ï¼?æ??建ä¸å?ç±»å??æ?¯è§?ã?ä¸å?空é?´å??辨ç??ã?ä¸å?波段ç»?å?ç??è®ç»?å½±å?é??å??æµ?è¯?å½±å?é??ï¼?å ¶æ¬¡å?ºäº?ä¸å?è®ç»?æ ·æ?¬é??å??æµ?è¯?é??ï¼?è¯?ä¼°äº?DeepLabv3+ç®?æ³?å?¨ä¸å?æ? å?µä¸?ç??äº?æ£?æµ?精度ï¼?并ä¸?Fmaskç®?æ³?ä½?对æ¯?å??æ?ã??ç ?究ç»?æ??表æ??ï¼?ï¼?1ï¼?使ç?¨å ¨æ··å?æ?¯è§?ç±»å??ç??è®ç»?é??è®ç»?å?ºæ¥ç??äº?æ£?æµ?ç½?ç»?å?¨æ?»ä½?æ£?æµ?精度ï¼?92.81%ï¼?ä¸?稳å®?度ï¼?æ ?å??å·®12.08%ï¼?ä¸?é?½ä¼?äº?使ç?¨å?ä¸?æ?¯è§?ç±»å??ç??è®ç»?é??è®ç»?å¾?å?°ç??äº?æ£?æµ?ç½?ç»?ï¼?ä¹?ä¼?äº?Fmaskç??æ?»ä½?精度ï¼?88.75%ï¼?ä¸?稳å®?度ï¼?æ ?å??å·®17.34%ï¼?ï¼?说æ??å?¨æ??建深度å¦ä¹ ç®?æ³?ç??è®ç»?é??æ?¶ï¼?åº?该尽å¯è?½å? å«å¤?ç±»å??ç??å?°è¡¨æ?¯è§?ï¼?ï¼?2ï¼?å°?å ¨æ··å?æ?¯è§?è®ç»?é??ä¸å??é?¤ä¸?ç±»æ?¯è§?ç??æ ·æ?¬ï¼?å?°/é?ªæ?¯è§?é?¤å¤?ï¼?æ??建ç??â??æ··å?-1â?è®ç»?é??ä¸?å ¨æ··å?æ?¯è§?è®ç»?é??è®ç»?ç??DeepLabv3+ç½?ç»?ç??äº?æ£?æµ?精度ä¹?ç?¸å·®ä¸å¤§ï¼?说æ??ç?°æ??è®ç»?æ ·æ?¬é??å·²å ·å¤?è¾?强ç??æ?¯è§?æ³?å??è?½å??ï¼?ï¼?3ï¼?å?ºäº?30 m空é?´å??辨ç??ç??å ¨æ··å?æ?¯è§?è®ç»?æ ·æ?¬é??è®ç»?å¾?å?°ç??DeepLabv3+äº?æ£?æµ?ç½?ç»?å?¨ä¸å?å??辨ç??ï¼?30 mã?60 mã?120 mã?240 mï¼?ç??æµ?è¯?é??ä¸?äº?æ£?æµ?精度差å¼?ä¸å¤§ï¼?é?½å?å¾?è¾?好ç??æ??æ??ï¼?说æ??DeepLabv3+è?½å¤?æ³?å??åº?ç?¨äº?ä¸å?空é?´å??辨ç??ç??é¥æ??å½±å?ï¼?ç?¸åFmaskç?´æ?¥åº?ç?¨äº?ä½?å??辨ç??å½±å?æ?¶ç²¾åº¦æ??æ?¾ä¸?é?ï¼?ï¼?4ï¼?DeepLabv3+è?½å å??è?ªé??åº?ä¸å?波段ç??ä¿¡æ¯ç?¨äº?äº?æ£?æµ?ï¼?æ?»ä½?æ¥è¯´æ?´å¤?ç??å ?谱波段è¾?å ¥è?½å¤?æé«?DeepLabv3+ç??äº?æ£?æµ?ç??精度å??稳å®?度ï¼?å ¶ä¸ç?波红å¤?波段对äº?DeepLabv3+å?ºå??å?°/é?ªä¸?äº?å ·æ??é?è¦ä»·å?¼ï¼?è??ç?红å¤?波段对DeepLabv3+äº?æ£?æµ?ç½?ç»?ç??æ?§è?½æå?å¾?å¾®å°ã??以ä¸?ç»?æ??说æ??å?©ç?¨ç?°æ??æ?°æ®é??â??L8 Biomeâ?è®ç»?ç??DeepLabv3+äº?æ£?æµ?ç½?ç»?è?½å¤?é??ç?¨äº?å¤?ç§ç±»å??ç??é¥æ??å½±å?ï¼?并ä¼?äº?Fmaskç®?æ³?ã??
منابع مشابه
application of upfc based on svpwm for power quality improvement
در سالهای اخیر،اختلالات کیفیت توان مهمترین موضوع می باشد که محققان زیادی را برای پیدا کردن راه حلی برای حل آن علاقه مند ساخته است.امروزه کیفیت توان در سیستم قدرت برای مراکز صنعتی،تجاری وکاربردهای بیمارستانی مسئله مهمی می باشد.مشکل ولتاژمثل شرایط افت ولتاژواضافه جریان ناشی از اتصال کوتاه مدار یا وقوع خطا در سیستم بیشتر مورد توجه می باشد. برای مطالعه افت ولتاژ واضافه جریان،محققان زیادی کار کرده ...
15 صفحه اولEdge detection for Very High Resolution satellite imagery based on Cellular Neural Network
In the context of Very High Resolution (VHR) satellite imagery, edge detection is one of the most important and difficult steps in image processing and pattern recognition. This paper presents the use of Cellular Neural Network (CNN) in edge detection and shows its capacity to locate and identify discontinuities in the gray levels of the objects that appear in VHR image. The results show that t...
متن کاملSatellite Imagery Classification Based on Deep Convolution Network
Satellite imagery classification is a challenging problem with many practical applications. In this paper, we designed a deep convolution neural network (DCNN) to classify the satellite imagery. The contributions of this paper are twofold — First, to cope with the large-scale variance in the satellite image, we introduced the inception module, which has multiple filters with different size at t...
متن کاملAutomatic 3d Change Detection Based on Optical Satellite Stereo Imagery
When monitoring urban areas from space, change detection based on satellite images is one of the most heavily investigated topics. In the case of monitoring change in 2D, one major shortcoming consists in the lack of height change detection. Thereby only changes related to reflectance values or local textures changes can be detected. However, changes in the vertical direction are completely ign...
متن کاملAssessment Methodology for Anomaly-Based Intrusion Detection in Cloud Computing
Cloud computing has become an attractive target for attackers as the mainstream technologies in the cloud, such as the virtualization and multitenancy, permit multiple users to utilize the same physical resource, thereby posing the so-called problem of internal facing security. Moreover, the traditional network-based intrusion detection systems (IDSs) are ineffective to be deployed in the cloud...
متن کاملذخیره 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: 2021
ISSN: ['1007-4619', '2095-9494']
DOI: https://doi.org/10.11834/jrs.20210061