• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) Ç×°ø ¿µ»ó ³» ¿ÀºêÁ§Æ®ÀÇ Á¤È®ÇÑ ¸éÀû ÃøÁ¤À» À§ÇÑ ¿Ü°û¼± °ËÃâ ¹æ¹ý
¿µ¹®Á¦¸ñ(English Title) An Edge Detection Method for Accurate Area Measurement of Objects in an Aerial Image
ÀúÀÚ(Author) ÇÑÀ±¼º   ±è¼öÁ¾   ±è¼®Áø   À̵¿¿í   Àü¼öºó   ¼­µ¿¸¸   Yunsung Han   Sujong Kim   Seokjin Kim   Dongwook Lee   Soobin Jeon   Dongmahn Seo  
¿ø¹®¼ö·Ïó(Citation) VOL 27 NO. 06 PP. 0282 ~ 0288 (2021. 06)
Çѱ۳»¿ë
(Korean Abstract)
ÃÖ±Ù µå·Ð »ê¾÷ÀÇ ¹ß´Þ¿¡ µû¶ó µå·Ð¿¡ °üÇÑ °ü½ÉÀÌ Ä¿Áö°í ÀÖÀ¸¸ç ±×·Î ÀÎÇØ, µå·ÐÀ¸·Î ÃÔ¿µÇÑ Ç×°ø ¿µ»ó ³» ¹°Ã¼¸¦ °ËÃâÇÏ°í ÆľÇÇÏ´Â ¿¬±¸°¡ È°¹ßÈ÷ ÁøÇà ÁßÀÌ´Ù. ±× Áß, µµ·Î¿¡¼­ ¿îÀüÀÚ¸¦ À§ÇùÇÏ´Â Æ÷ƮȦ, Å©·¢ÀÇ ¸éÀû°ú À§Çèµµ ÃøÁ¤À» ÀÚµ¿È­Çϱâ À§ÇØ Ç×°ø ¿µ»ó ³» ¹°Ã¼¸¦ °ËÃâÇÏ°í ÇØ´ç ¹°Ã¼ÀÇ ¿µ¿ªÀ» ÃøÁ¤ÇÏ´Â ¿¬±¸°¡ ÁøÇàµÇ°í ÀÖ´Ù. µå·Ð¿¡¼­ ÃÔ¿µÇÑ ¿µ»óÀÇ °íµµÁ¤º¸¿Í ¿µ»óÀÇ ¸ÞŸµ¥ÀÌÅ͸¦ ÀÌ¿ëÇÏ¿© Çȼ¿ ´ç ´ÜÀ§ ¸éÀûÀ» ¾òÀ» ¼ö ÀÖ´Ù. À̸¦ ÅëÇØ À§Çè¹° ¸éÀû °è»êÀÌ °¡´ÉÇÏ¸ç ´õ¿í Á¤È®ÇÑ À§Çèµµ ÆÇ´ÜÀ» À§ÇØ Á¤È®ÇÑ ¸éÀû °è»êÀÌ ÇÊ¿äÇÏ´Ù. º» ³í¹®¿¡¼­´Â Mask R-CNNÀ» ÅëÇØ ¹°Ã¼ÀÇ ¸¶½ºÅ©¸¦ ÃßÃâÇÏ°í ¿Ü°û¼± °ËÃâ ¾Ë°í¸®ÁòÀ» ÀÌ¿ëÇÏ¿© ¸¶½ºÅ© ³» ¿Ü°û¼±À» °ËÃâÇÑ ÈÄ, ÃÖ±ÙÁ¢ ÀÌ¿ô ¾Ë°í¸®ÁòÀ» ÅëÇØ ¿¬°áµÈ ¿Ü°û¼±À» ¸¸µé¾î Á¤È®ÇÑ ¹°Ã¼ÀÇ ¿µ¿ªÀ» ÃßÃâÇÏ´Â ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ±âÁ¸ ¿Ü°û¼± °ËÃâ ¾Ë°í¸®Áò ¹× ÃÖ±ÙÁ¢ ÀÌ¿ô ¾Ë°í¸®ÁòÀ» Æò°¡ÇÏ¿© Mask R-CNNÀ¸·Î °ËÃâµÈ ¸¶½ºÅ© ¿µ¿ªº¸´Ù Á¦¾È¹æ¹ýÀÌ Á¤È®ÇÑ ¸éÀûÀ» ÃßÃâÇÏ´Â °ÍÀ» È®ÀÎÇÏ¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Recently, in response to the development of the drone industry, there has been a growing interest in drones and research to understand by detecting the object in the video obtained by a drone has been actively promoted. Among these researches, studies are underway to detect the objects in aerial images and measure the areas of those objects in order to automate measuring the area and risk of potholes and cracks that threaten the driver on the road. In the case of images taken at high altitudes, such as those taken by drones, the exact area measurement is an important indicator of risk because the unit area per pixel is wide and it affects the calculation of the area of dangerous goods. In this paper, we propose the method for extracting the mask of the object through Mask R-CNN and detecting the outline within the mask by using the contour detection algorithm, and then create a connected contour through the nearest neighbor algorithm to extract the area of the correct object. The existing border detection algorithm and the nearest neighbor algorithm were evaluated to confirm that the method proposed in this paper extracts the correct area than the mask area detected by Mask R-CNN.
Å°¿öµå(Keyword) ¿Ü°û¼± °ËÃâ   ¿Ü°û¼± ¿¬°á   Ç×°ø ¿µ»ó   ÃÖ±ÙÁ¢ ÀÌ¿ô ¾Ë°í¸®Áò   °´Ã¼ºÐÇÒ µö·¯´×   edge detection   edge connect   aerial image   nearest neighbor   instance segmentation deep learning  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå