浠庨害鐢板浘鍍忎腑妫€娴嬪拰琛ㄥ緛绌楀瓙瀵逛簬绮剧‘鍐滀笟鐨勫皬楹︾敓闀跨洃娴嬭嚦鍏抽噸瑕併€傞殢鐫€鍚勭鎶€鏈殑鍙戝睍锛屽熀浜庢繁搴﹀涔犵殑鏂规硶宸茬粡鏄捐憲鍦版彁楂樹簡灏忛害绌楃殑妫€娴嬫€ц兘銆傜劧鑰岋紝鐢变簬鐢伴棿灏忛害楂樼閬尅鍜屽鏉傝儗鏅細瀵艰嚧閿欐鍜屾紡妫€鐨勯棶棰橈紝鍦ㄦ棤浜烘満鍥惧儚涓娴嬪皬鐨勫拰閲嶅彔鐨勫皬楹︾浠嶅叿鏈夋寫鎴樻€с€傛湰鏂囨彁鍑轰簡涓€绉嶇敤浜庡畾鍚戝皬灏哄灏忛害绌楁娴嬬殑娣卞害瀛︿範鏂规硶锛圤SWSDet锛夈€備笌缁忓吀鐨勫皬楹︾绮掓娴嬫柟娉曚笉鍚岋紝OSWSDet閫氳繃闆嗘垚鍦嗗舰骞虫粦鏍囩锛圕SL锛夊拰寰昂搴︽娴嬪眰锛屽皢灏忛害绌楃矑鐨勬柟鍚戝紩鍏OLO妗嗘灦銆傝繖浜涙敼杩涙彁楂樹簡妫€娴嬪皬灏哄楹︾鐨勮兘鍔涳紝闃叉浜嗛害绌楁娴嬮敊璇€傚疄楠岀粨鏋滆〃鏄庯紝OSWSDet浼樹簬缁忓吀鐨勯害绌楁娴嬫柟娉曪紝骞冲潎绮惧害锛圓P锛変负90.5%銆侽SWSDet鑳藉鍑嗙‘鍦版娴嬪嚭鍏锋湁澶嶆潅鐢伴棿鑳屾櫙鐨勬棤浜烘満鍥惧儚涓殑楹︾锛屼负浠婂悗鐢伴棿灏忛害琛ㄥ瀷鐩戞祴鎻愪緵浜嗘妧鏈弬鑰冦€?nbsp;
鍥? 瀹為獙鍦扮偣鍜屾棤浜烘満鐨勯害绌楀浘鍍忋€?a)娓呮櫚鐨勫浘鍍忥紝(b)妯＄硦鐨勫浘鍍忥紝(c)浜哄伐鏍囨敞鐨勫浘鍍忋€侟/div>
鍥? OSWSDet 鐨勬€讳綋妗嗘灦銆傜孩鑹叉爣绛炬槸OSWSDet涓彁鍑虹殑鏀硅繘锛屽寘鎷 CSL锛堝渾褰㈠钩婊戞爣绛撅級銆佸井灏哄害妫€娴嬪眰銆佸熀浜 CIoU锛圕omplete Intersection over Union锛夊拰 RNMS锛圧otation Non-Maximum Suppression锛夌殑浼樺寲瀹氫綅鎹熷け銆侟/div>
鍥? 鎹寮猴細锛坅锛夊師濮嬪浘鍍忥紝锛坆锛夋棆杞?0°锛岋紙c锛夋棆杞?80°锛岋紙d锛夋棆杞?70°锛岋紙e锛夊瀭鐩寸炕杞紝锛坒锛夋按骞崇炕杞紝锛坓锛夊拰锛坔锛変寒搴﹀钩琛°€侟/div>
鍥? 涓や釜瀹氬悜妫€娴嬫鐨 WIoU 璁＄畻銆傚乏鍥炬樉绀轰簡灏忛害绌楃殑瀹氬悜妫€娴嬫锛屽彸杈圭殑6骞呭浘鍍忔樉绀轰簡涓や釜閲嶅彔妗嗕箣闂撮噸鍙犲尯鍩熺殑褰㈢姸銆傛牴鎹舰鐘剁殑涓嶅悓锛屼袱涓噸鍙犳鏈?绉嶄笉鍚岀被鍨嬬殑榛勮壊鐩镐氦鍥撅紝鍙互鐢ㄤ笁瑙掑舰鐨勯潰绉潵璁＄畻WIoU锛氾紙a锛夌浉浜ゅ浘鏄笁瑙掑舰锛岋紙b锛夌浉浜ゅ浘鏄洓杈瑰舰锛岋紙c锛夌浉浜ゅ浘鏄簲杈瑰舰锛岋紙d锛夌浉浜ゅ浘鏄叚杈瑰舰锛岋紙e锛夌浉浜ゅ浘鏄竷杈瑰舰锛岋紙f锛夌浉浜ゅ浘鏄叓杈瑰舰銆侟/div>
鍥? 鎵€鎻愭柟娉曞拰鍏朵粬鏈€鍏堣繘鏂规硶鐨勫皬楹︾瀛愭娴嬬殑绮剧‘鎬у拰鍙洖鐜囨洸绾裤€侟/div>
鍥?鍥惧儚涓婃爣娉ㄧ殑鐪熷疄绌楀瓙鏁伴噺涓嶧aster RCNN銆丼SD銆乻tandard YOLOv5銆乻tandard YOLOv5with micro-sclae layer銆乻tandard YOLOv5with micro-sclae layer and CIoU浠ュ強OSWSDet浼拌鐨勭瀛愭暟閲忕殑姣旇緝銆侟/div>
鍥? 瑙掑害绛夌骇鐨勫皬楹︾鏁颁粠0鍒?79锛屽叿鏈?00涓互涓婂皬楹︾鐨勭被鏍囨湁鐗瑰畾缂栧彿銆侟/div>
鍥?浣跨敤OSWSDet锛坅锛夊拰鏍囧噯YOLOv5锛坆锛夌殑妫€娴嬬粨鏋溿€傜孩鑹茶竟妗嗕唬琛ㄦ娴嬬粨鏋溿€侟/div>
鍥?浣跨敤CSL锛坅锛夊拰鏃燙SL锛坆锛夎搴﹀垎绫荤殑妯″瀷妫€娴嬬粨鏋溿€傜孩鑹茶竟妗嗚〃绀烘娴嬬粨鏋溿€侟/div>
鍥?0 涓嶅悓澶у皬鐨勭瀛愮殑鏁伴噺鍒嗗竷銆侟/div>
鍥?1鍦ㄨ缁冭繃绋嬩腑锛岀敤寰昂搴︽娴嬪眰锛坅锛夈€佸皬灏哄害妫€娴嬪眰锛坆锛夈€佷腑灏哄害妫€娴嬪眰锛坈锛夊拰澶у昂搴︽娴嬪眰锛坉锛夊疄鐜扮殑姝ｆ牱鏈竟鐣屾鍜岄害绌楄竟鐣屾銆傛鏍锋湰杈圭晫妗嗭紙娴呰摑鑹诧級銆傚皬楹︾鐨勮竟鐣屾锛堢矇绾㈣壊锛夈€侟/div>
鍥?2浣跨敤CIoU杩涜瀹氫綅鎹熷け鐨勬ā鍨嬬殑妫€娴嬬粨鏋滐紙a锛夊拰浣跨敤IoU杩涜瀹氫綅鎹熷け鐨勬娴嬬粨鏋滐紙b锛夈€傜孩鑹茶竟妗嗕唬琛ㄦ娴嬬粨鏋溿€
鏉ユ簮
Zhao J, Yan J, Xue T, et al. A deep learning method for oriented and small wheat spike detection (OSWSDet) in UAV images[J]. Computers and Electronics in Agriculture, 2022, 198: 107087.https://doi.org/10.1016/j.compag.2022.107087