2019³â Ãß°èÇмú´ëȸ
Current Result Document :
ÇѱÛÁ¦¸ñ(Korean Title) |
Á¾ÀÚƯ¼º Á¤º¸ÃßÃâ ÇÁ·Î±×·¥ °³¹ß |
¿µ¹®Á¦¸ñ(English Title) |
Development of Seed Traits Information Extraction Program |
ÀúÀÚ(Author) |
¹éÁ¤È£
ÀÌÀº°æ
±è³âÈñ
ÀÌÈ«¼®
±è¼Û¸²
ÃÖÀÎÂù
ÁöÇö¼Ò
±è°æȯ
JeongHo BAEK
Eungyeong LEE
Nyunhee KIM
Hongseok LEE
Song Lim KIM
Inchan CHOI
Hyenso JI
Kyung-Hwan KIM
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 23 NO. 02 PP. 0569 ~ 0571 (2019. 10) |
Çѱ۳»¿ë (Korean Abstract) |
½Ä¹°ÀÇ Á¾ÀÚ´Â ½Ä·®, ¼ÒÀç, Á¾ÀÇ À¯Áö ¹× ¹ø½Ä µî ´Ù¾çÇÑ È°¿ëÀ¸·Î ³ó¾÷¿¡¼ ¸Å¿ì Áß¿äÇÑ Àǹ̸¦ °¡Áø´Ù. À°¾È°ú ¾Æ³¯·Î±×Àû ¹æ¹ýÀ» ÅëÇØ ¿ì¼öÇÑ Á¾ÀÚ¸¦ ¼±º°Çϱâ À§ÇÑ ¿¬±¸°¡ Áö¼ÓÀûÀ¸·Î ÁøÇàµÇ¾î ¿ÔÀ¸¸ç ÃÖ±Ù ¿µ»óÀ» ÀÌ¿ëÇÏ¿© Á¤¹ÐÇÏ°í ´ë·®À¸·Î Ư¼ºÀ» È®ÀÎÇÏ´Â ¿¬±¸µéÀÌ È°¹ßÈ÷ ÁøÇàµÇ°í ÀÖ´Ù. Á¾ÀÚ´Â ´Ù¾çÇÑ Å©±â, ¸ð¾ç µî ÇüÅÂÀûÀΠƯ¡°ú Á¾ÀÚÀÇ »ö±ò(Á¾ÇÇ»ö), Á¾ÀÚ³» Æ÷ÇÔµÈ ¿µ¾ç¼Ò µîÀ¸·Î °¢°¢ Á¾°ú Ư¼ºÀÌ ±¸º°µÇ³ª »ç¶÷ÀÌ ¹Ì¹ÌÇÑ Â÷À̸¦ ±¸º°ÇÏ´Â °ÍÀÌ ½±Áö ¾Ê°í ´ë·®À¸·Î Ư¼º ºÐ¼®ÇÏ´Â °ÍÀº ºÒ°¡´É¿¡ °¡±õ´Ù. º» ¿¬±¸¿¡¼´Â Äá°ú º Á¾ÀÚ 800Ç°Á¾ 160,000¸³¿¡ ´ëÇØ µðÁöÅÐÄ«¸Þ¶ó¸¦ ÀÌ¿ëÇÏ¿© ÃÔ¿µÇÏ¿´°í, YUVÀÇ »öÂ÷Á¤º¸¸¦ È°¿ëÇÏ¿© ¹è°æÀ¸·ÎºÎÅÍ Á¾ÀÚ°´Ã¼ ºÐ¸®ÈÄ area, width, height, circulation, roundness µî 11°³ÀÇ ÇüÅÂÁ¤º¸¿Í RGB »ö»ó È÷½ºÅä±×·¥À¸·Î Ư¼ºÀ» ÃßÃâÇÏ´Â ÇÁ·Î±×·¥À» °³¹ßÇÏ¿´´Ù. ÇÁ·Î±×·¥À» Åë ÇØ ÃßÃâÇÑ Æ¯¼º°ª°ú ½ÇÃø°ª°úÀÇ »ó°ü°ü°è¸¦ ºÐ¼®ÇÑ °á°ú 0.95ÀÌ»óÀ¸·Î ³ôÀº ¿¬°ü¼ºÀ» È®ÀÎÇÏ¿´´Ù. ÀÌ·¸ °Ô ÃßÃâÇÑ Á¤º¸µéÀº Á¾ÀÚƯ¼º ºÐ·ù¸¦ ÅëÇØ ¿ì¼öÁ¾ÀÚ ¼±¹ßÀ̳ª À¯ÀüüÁ¤º¸¿Í Ç¥ÇöÇüÁú ¿¬°è¿¡ µû¸¥ À¯ÀüÀÚ ±â´É È®ÀÎ µî °£Æí Áø´Ü µî¿¡ È°¿ëÇÒ ¼ö ÀÖ¾î À°Á¾´Ü°è¿¡¼ ½Ã°£°ú ³ë·ÂÀ» ´ÜÃàÇϴµ¥ Å« ±â¿©¸¦ ÇÒ ¼ö ÀÖ´Ù. ÇâÈÄ ÃßÃâµÈ Á¾ÀÚƯ¼º Á¤º¸·Î ±â´É¼º ¹°ÁúÀÌ Æ÷ÇÔµÈ °íºÎ°¡ °¡Ä¡¸¦ °®´Â ½ÅÇ°Á¾ ÀÛ¹°À» À°¼ºÇϴµ¥ È°¿ëÇÒ ¿¹Á¤ÀÌ´Ù. |
¿µ¹®³»¿ë (English Abstract) |
Seeds of plants are very important meaning with various uses such as food, materials, maintenance and reproduction. Seeds have been steadily researched since the past, and recently, studies are being actively conducted to measure seed traits precisely in large scale using images. Seeds can also represent the traits of their plants by having morphological features, colors and nutrients including various sizes and shapes. However, it is not easy for a person to distinguish small differences and mass analysis is almost impossible. In this paper, we developed a program to extract the shape and color informations of seeds with about 160,000 seeds of soybean and rice using images. Photographs of seeds were taken using a digital camera, and the background was removed using the color difference information of YUV. The extracted seed trait informations include 11 shape-related informations such as area, width, height, circulation, roundness of seed and RGB color histogram. In addition, high correlations of higher than 0.95 were confirmed by comparing seed trait values measured using informations extracted from the images with the values from manual measurement of seed traits. |
Å°¿öµå(Keyword) |
Á¾ÀÚ
¿µ»ó
Ư¼ºÁ¤º¸
ÇÁ·Î±×·¥
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|