• 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 ³í¹®Áö D : µ¥ÀÌŸº£À̽º

Á¤º¸°úÇÐȸ ³í¹®Áö D : µ¥ÀÌŸº£À̽º

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) Ãʺб¤¿µ»óÀÇ ¹°ÁúŽÁö¸¦ À§ÇÑ ÇǶó¹Ìµå ±â¹ý ±â¹Ý »öÀÎ ¹× K-ÃÖ±ÙÁ¢ ÁúÀÇ ¾Ë°í¸®Áò
¿µ¹®Á¦¸ñ(English Title) An Indexing and K-NN Query Algorithm based on Pyramid-technique for Target Detection of Hyperspectral Image
ÀúÀÚ(Author) À¯Àçȯ   ½ÅÁ¤ÀÏ   ±è´öȯ   Jae-Hwan Yu   Jung-il Shin   Deok-Hwan Kim  
¿ø¹®¼ö·Ïó(Citation) VOL 39 NO. 03 PP. 0180 ~ 0192 (2012. 06)
Çѱ۳»¿ë
(Korean Abstract)
Ãʺб¤¿µ»óÀº ´ÙÁß ºÐ±¤ ¿µ»óº¸´Ù ¸¹Àº ¹êµå¸¦ ÅëÇÏ¿© ³ÐÀº ¹üÀ§ÀÇ ÆÄÀå ¿µ¿ª¿¡ ´ëÇÑ ¹Ý»çÀ²À» ´ã°í ÀÖ´Â °íÂ÷¿ø µ¥ÀÌÅÍÀÌ´Ù. ÀÌ·¯ÇÑ °íÂ÷¿ø µ¥ÀÌÅÍÀÇ »öÀÎ ±¸ÃàÀ» ÇÒ ¶§¿¡´Â Â÷¿øÀÇ ÀúÁÖ ¶§¹®¿¡ ¼º´ÉÀÌ ±Þ°ÝÈ÷ ÀúÇϵȴÙ. ¶ÇÇÑ ÁúÀÇ Ã³¸® °úÁ¤¿¡¼­ ƯÁ¤ ¹êµåÀÇ ¹Ý»çÀ² °ªÀÌ ºñ½ÁÇÑ ´Ù¸¥ ¹°ÁúÀÌ Èĺ¸ÀÚ·Î °Ë»öµÉ ¼öµµ ÀÖ´Ù. º» ³í¹®¿¡¼­´Â Â÷¿øÀÇ ÀúÁÖ ¹®Á¦¸¦ ÇØ°áÇϱâ À§ÇÏ¿© ÇǶó¹Ìµå ±â¹ýÀ» ÀÌ¿ëÇÑ Ãʺб¤¿µ»ó »öÀÎÀ» ±¸ÃàÇÏ¿´´Ù. ¶ÇÇÑ ºÐ±¤ ¶óÀ̺귯¸®¿¡¼­ ´ë»ó ¹°ÁúÀ» ÃßÃâÇϱâ À§ÇØ ¹ÌºÐ ºÐ±¤¹ýÀ» ÀÌ¿ëÇÑ Àüó¸®¿Í ÇǶó¹Ìµå ±â¹ý ±â¹ÝÀÇ K-ÃÖ±ÙÁ¢ ÁúÀǸ¦ Á¦¾È ÇÏ¿´´Ù. ÀÌ¿Í ´õºÒ¾î ÁúÀÇ Ã³¸® °úÁ¤¿¡¼­ °ÅÁþ °ªÀ» °É·¯³»±â À§ÇÏ¿© ÀûºÐ ºÐ±¤ Á¤ÇÕ ÇÊÅ͸¦ Ãß°¡ÇÏ¿´´Ù. ½ÇÇè °á°ú Á¦¾ÈÇÑ ¹æ¹ýÀº k=10 ÀÏ ¶§ ¼øÂ÷Àû Á¢±Ù ¹æ¹ý¿¡ ºñÇØ 4%ÀÇ ½Ã°£ ¹Û¿¡ °É¸®Áö ¾Ê¾Ò´Ù. ¶ÇÇÑ 84%ÀÇ Á¤Å½ÁöÀ²°ú 16%ÀÇ ¿ÀŽÁöÀ²À» º¸¿´´Ù.
¿µ¹®³»¿ë
(English Abstract)
Hyperspectral image has large band than multispectral image. Thus, it is high dimensional data of including reflectance for wide range of wave length. Indexing the high dimensional data may decrease the performance due to curse of dimensionality problem. Besides, query processing may retrieval other material whose reflectance is similar to specific bandwidth. In this paper, we propose hyperspectral image indexing using pyramid technique to resolve the curse of dimensionality problem and preprocessing using derivative analysis and K-NN query based on pyramid index to retrieval target material in the spectral library. Besides, integral spectral matching filter was added in the query processing phase to decrease false hit ratio. Experiment results show that the retrieval time of the proposed method is 96% less then that of sequential access when k=10, and its recall and false detection ratio are 84%, and 16%, respectively.
Å°¿öµå(Keyword) Ãʺб¤¿µ»ó   ºÐ±¤ ¶óÀ̺귯¸®   ÆĶó¹Ìµå »öÀΠ  ÀûºÐ ºÐ±¤ Á¤ÇÕ ÇÊÅÍ   ¹ÌºÐ ºÐ±¤¹ý   ¹°Áú ŽÁö   Hyperspectral image   Spectral library   Pyramid indexing   Integral spectral matching filter   Derivative spectral analysis   Target detection  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå