文档介绍：ѢϡৠሎᑺঠⳂ㾚㾝ⱘ AGV ᇐᓩᮍ⊩ⷨお
ཷಷϬēვॿէ
ϰफᄺ㋏㒳Ꮉⷨお᠔∳㢣फҀ
ࡂēწճંঠԅݮဟγಭຬಭιԅܬᨬ㽕˖นॴඔಖ$*9 ԅᄦટഃ଼ēٗۚԙॆဈ
ԅނࡂۤϰܬၮ 9*$ ڶೊࢄӽูඕēඔѻॴྡྷᄵϢලѤէԅഀંೊࢄ$*9 ӽֺ֥dඹ 9*$
ԅݮ҄౨ē࠙ߋտѤէϏຸēဈ 6REHOىדֺ֥ēၽᗰᄔىדॴ݇էᗰᄔܤࠩۤည؟ēד
юဈ 5REHUWV സᆐࠩ໌ѤէΨၘЉēݦॴԁނസᆐۤӖѤէϏ੦͐ݦԁাԅ͟Ψၘēω
ඔ٢ॴ॥ۚԅᄆћdලನēॆဈѤէϢέඋჸέܰď6FDOH,QYDULDQW۔াԅ࠺௲ՇสΨၘēน$*9 ӽ
ēճ$*9 ࠩԁাؘ࠭ԅა̘דۦHDWXUH7UDQVIRUPē6,)7Đസ֥ēճ०ᄵϢලѤէԅࠩక(
ࠩЉēωٓѻॴ०ᄵϢලѤէԅତଅֺ֥ēྻಬຣഀંೊࢄԅೊЕݦd
݇䬂䆡ġϢලѤէēഀંೊࢄē$*9ēೊࢄӽēѤէϢέඋჸέܰ
Ёߚ㉏ো˖TP ᭛⤂ᷛ䆚ⷕ˖A
AGV Guidance Methods based on Various Scale Binocular Vision
YANG Shi-cai, ZHAO Lin-du
(Institute of Systems Engineering, Southeast University, Nanjing 210096, China)
Abstract: To improve the intelligent level of AGV and make better use of environment information, a vision guidance method of
various scale binocular vision using landmarks is presented. First, to get a better threshold value, the gray threshold segmentation
method is improved and optimized based on the analysis on the environment and the image. Second, drawing on the ideas of
multi-scale filtering, the Sobel operator and large-scale filter template are used to obtain the robust edge of the road, and then Roberts
operator is used to carry out the small- scale edge detection to obtain a precise positioning of the edge of the road, which could
provides a good support for the guidance of AGV. Third, SIFT(Scale Invariant Featur