isiZulu
Izinhlelo Zokubona Ezisebenzini Zemikhumbi Ezizenzakalelayo: Umshayeli Ongabonakali Wokusebenza Kahle Kwezohwebo Lomhlaba
Kwadalwa ngo 01.28
Ukuhweba komhlaba wonke kuncike ekuhambeni okungenamihawu kweziqukathi, futhi iziteshi zeziqukathi ezizenzakalelayo ziye zavela njengomgogodla walezi zinhlelo zezinto zokuhambisa. Ezingxenyeni ezibalulekile zalezi ziteshi zobuchwepheshe obuphezulu kukhona ingxenye ebalulekile kodwa evame ukunganakwa: izinhlelo zokubona. Ngaphezu kwamakhamera alula, ubuchwepheshe bokubona banamuhla—obutshazwe yi-AI, ukufunda komshini, nokuthwebula okuthuthukile—buyayiguqula indlela ama-crane azenzakalelayo asebenza ngayo, kuxazululwa ukungasebenzi kahle okuhlala njalo nezingozi zokuphepha. Lesi sihloko sihluza izicelo ezintsha, izinzuzo ezibonakalayo, kanye nezitayela zesikhathi esizayo ze izinhlelo zokubona ezitshala izikebhe zamaloli ezizenzakalelayo, zembula ukuthi kungani ziba yinto eyimpoqo ezitshaleni ezifuna ukuhlala ziqhudelana emakethe yomhlaba.
Izinkinga Ezixazululwa Izinhlelo Zokubona: Ngaphezu Kwemikhawulo Eyenziwa Ngabantu
Imisebenzi ejwayelekile yamachweba ezithwali, ngisho nalezo ezinezingxenye ezizenzakalelayo, ibilokhu ilwa nezinkinga ezibangelwa imikhawulo yabantu nobuchwepheshe obudala. Ukusebenza kwezinsimbi ngesandla kudinga abashayeli abanekhono eliphezulu ukuze bahambise imishini emikhulu ngokunemba, bahambise izikhiye zokubopha ezindaweni zokubopha izithwali—izimbobo ezincane kunezinqindi—ezingxenyeni eziphakeme. Le nqubo, eyaziwa ngokuthi "ukuhambisa izikhiye," ayisindi isikhathi kuphela kodwa futhi ingaba namaphutha ngenxa yokukhathala, ukungaboni kahle, nezimo zezulu ezimbi. Ngaphezu kwalokho, ukulandelela izinombolo zezithwali nezimo ngesandla kuholela ekubambezelekeni, ukungabi khona kweqiniso kwedatha, nokwanda kwezindleko zabasebenzi.
Lezi zinselelo ziba nzima kakhulu ngenxa yesidingo esikhulayo sezikhathi ezifushane zokuqedela imisebenzi. Izimboni zomhlaba wonke zidinga ukuthi iziteshi zezikebhe ziphathe izikebhe ezinkulu (ezikwazi ukuthwala ama-TEU angaphezu kuka-24,000) futhi ziphumelele iziqukathi eziningi kunangaphambili. Ngokwedatha yezimboni, imisebenzi yendabuko eyenziwa ngesandla ingandisa izikhathi zokujikeleza amaloli zibe ngemizuzu engu-56 noma ngaphezulu, kanti amaphutha omuntu ekusingatheni iziqukathi abangela ukulahleka kokusebenza okungu-15% nsuku zonke. Izinhlelo zokubona zixazulula lezi zinkinga ngqo, zivumela imishini yokukhomba ukuthi "ibone" futhi "iphendule" ngokunemba, isivinini, nokwethembeka okukhulu kunezisebenzi zabantu.
Ubuhlakani Bezobuchwepheshe Obunika Amandla Izinhlelo Zokubona Zesimanje Zesikebhe
Namuhla izinhlelo zokubona ezitshaleni zamaloli ezizenzakalelayo zihlanganisa ihadiwe ethuthukisiwe kanye nesofthiwe ehlakaniphile, eklanyelwe ukusebenza ezindaweni ezinzima kakhulu zesikebhe. Nansi ubuhlakani obubalulekile bezobuchwepheshe obuqhuba ukusebenza kwazo:
1. Amamodeli Okufunda Okujulile Anikwe Amandla yi-AI
Emoyeni wezinhlelo zesimanje zokubona, kukhona amamodeli e-AI ajulile aqeqeshwe ezithombeni zeziqukathi eziyizigidi. Amamodeli angakwazi ukukhomba ngokunemba izinombolo zeziqukathi, amakhodi e-ISO, iziqondiso zezicabha, ngisho nokulimala okuncane—ngisho nasezikhathini zokukhanya okuncane, imvula, noma inkungu. Ngokwesibonelo, uhlelo lwe-Huaming Vision lokubona izinombolo zeziqukathi ezisekhreynini lolwandle lufinyelela izinga lokubona elingu-98%+ ngokusebenzisa izithombe eziningi zemibala nokufunda okujulile, kucubungula idatha yesiqukathi esisodwa ngemizuzwana engu-0.3 nje. Lokhu kuyisinyathelo esikhulu kakhulu esivela ekurekhodweni ngesandla, okwakuthatha imizuzu emi-3 isiqukathi ngasinye futhi kwakuvame ukuba nephutha lomuntu.
Enye impumelelo ukusetshenziswa kwama-dataset ayimfihlo aqeqeshwe kusengaphambili kanye nokufundiswa kolwazi. Izinhlelo ezifana ne-AI Vision Large Model Container Quay Crane Roadside Automation System, eyakhiwe yi-Fujian Electronic Port Co., Ltd., zisebenzisa inethiwekhi yothisha (eqeqeshwe kusengaphambili kuma-dataset amakhulu) ukuhola imodeli yokukhiqiza. Lokhu kuvumela uhlelo ukuthi lucubungule imifudlana yesithombe yesikhathi sangempela ngokushesha, luhlonze iziqukathi nezakhiwo zamaloli ngokunemba okuphezulu. Umphumela yisikhathi sokuphendula samasekhondi angu-40 sokubekwa kwemoto yonke kanye nokubona imbobo yokukhiya—ngokushesha kakhulu kunezixazululo ezisekelwe ku-laser ezivamile ezithatha amasekhondi angu-3-5.
2. Ukuhlanganiswa Kwezinzwa Eziningi Nokubona Okubili
Ukuze kunqotshwe imikhawulo yezinhlelo zekhamera eyodwa, izixazululo zesimanje zokubona zihlanganisa amakhamera amaningi ezimboni asebenza kakhulu anama-laser radar, kudale uhlelo lwe-stereo vision. Lesi simo sihlinzeka ngokubeka izikhundla ze-3D, sibala izixhumanisi eziqondile ze-X, Y, Z nama-engeli okujikeleza ezikhuthini ngokunemba okulinganiselwa ku-centimeter. Ngokwesibonelo, isilawuli se-AI gantry crane esisekelwe ku-Ubuntu sisebenzisa inhlanganisela yamakhamera kanye ne-laser radar ukuskena izindawo zezikhuthini nezimoto zokuhamba, sivumela i-crane ukuthi ilungise ngokuzenzakalelayo isikhundla se-spreader kanye ne-engeli ukuze kube nokukhiya okunembayo.
Ukuqinisekiswa kokubuyekezwa kwezimbili-ikhamera kungenye into ebalulekile. Ngokusebenzisa amakhamera amabili ukuhlola idatha yomunye nomunye, izinhlelo zingasusa amaphutha abangelwa ukuhlanekaniswa kwekhamera noma ukufihla. Izindlela ezithuthukisiwe zokulungisa ngokuzenzakalelayo ukuhlanekaniswa zithuthukisa ukunemba ngokubala nokukhokhela ukuhlanekaniswa kwezibukweni zamehlo ngokusebenzisa ukubalwa kombono.
3. Ukubala Kwe-Edge ukuze Kuphendule Ngempela
I-Latency iyisitha esikhulu emisebenzini yokusebenza kwephothi okuzenzakalelayo, njengoba noma ukulibaziseka okuncane kungaholela ekuphahlazweni noma ekungahambisani. Izinhlelo zokubona manje zisebenzisa ukubala kwe-edge, zisebenzisa ama-algorithms ngqo kumakhompyutha ezimboni akhompuyutha ezikhepheni kunokuthembela kumaseva akude efu. Lokhu kunciphisa ukulibaziseka kokudluliswa kwedatha, kuvumela impendulo ye-millisecond kumkhankaso wokulawula we-crane.
I-Shore Crane Automatic Tally System, ngokwesibonelo, isebenzisa ikhompyutha ye-AI eyakhelwe ngaphakathi ukucubungula idatha ebonakalayo emaphethelweni. Lokhu kuyivumela ukuthi ikhiphe izici zesiqukathi (isibonelo, uhlobo, usayizi, umonakalo) futhi ithumele imiyalelo ohlelweni lwe-PLC lwe-crane ngesikhathi sangempela, iqeda ukulahleka kwekhwalithi nokubambezeleka okuhlotshaniswa nokufakwa kwevidiyo/ukukhipha amakhodi ezinhlelweni ezivamile ezisekelwe emafini.
Izicelo Zangempela: Ukuguqula Amachweba Emhlabeni Wonke
Izinhlelo zokubona azizona nje izinto ezintsha zemibono—sezivele zikhiqiza imiphumela ebonakalayo kumanye amachweba ashukumisayo emhlabeni. Ake sihlole izibonelo ezimbili ezivelele ezibonisa umthelela wazo:
Isifundo Sendawo 1: Izinhlelo Zokubona Ze-AI Ezitshala Izikebhe ZaseShayina NezaseNingizimu-mpumalanga Asia
Uhlelo Lwe-AI Vision Large Model Lwe-Fujian Electronic Port selufakwe eziteshini ezingaphezu kuka-30 emhlabeni wonke, okubalwa kuzo iShanghai Port, Qingdao Port, Xiamen Port, Singapore Port, kanye neSantos Port yaseBrazil. Uhlelo lusekela imisebenzi emithathu eyinhloko: imisebenzi ezansi komhlaba ezenzakalelayo, ukuqondisa izilayidi zangaphandle ze-yard crane, kanye nokuqondisa izilayidi zesibili. Emisebenzini ezansi komhlaba, luqondisa i-spreader endaweni enembayo ngokuthola izixhumanisi ezansi komhlaba, lulungise isikhala sokukhiya se-spreader kanye ne-engeli kusengaphambili ukuze kukhethwe iziqukathi ngokunembayo.
Imiphumela iyathakazelisa: izikhathi zokulinda amaloli zincishiswe kakhulu, futhi uhlelo lusize abaqhubi bamachweba ukuthi banciphise izindleko zabasebenzi kanye nesikhathi sokungasebenzi kwemishini. Ngokwenani lenkontileka edlula ama-RMB ayizigidi ezingu-25, kube umshayeli oyinhloko wokuguquka kwedijithali kwamachweba aseChina kanye nokuthuthukiswa kokuncintisana kwamazwe omhlaba.
Isifundo Sendawo 2: Izinhlelo Zokubona Ze-OCR Ze-Visy Eziteshini Zase-APM Gothenburg
I-APM Terminals Gothenburg, ephatha ama-TEU angaphezu kuka-800,000 ngonyaka, yasebenzisana ne-Visy ukufaka isixazululo esiphelele sezinhlelo zokubona. Isigaba 1 sifake izimbobo ze-OCR zamaloli nezesitimela, kuncishiswe isikhathi sokujikeleza kwamaloli kusuka kumaminithi angu-56 kuya kumaminithi angu-25 futhi kuncishiswe ukukhishwa kwekhabhoni ngamathani angu-688 ngo-2020. Isigaba 2 sengeze izinhlelo ze-OCR zamakhenane ezinekhamera ze-spreader ezibuka phezulu ezikheneni ezingu-7 ze-STS nezikheneni ezingu-2 ze-RMG, zafinyelela izinga lokubonwa kwe-ID yesiqukathi elingu-99%+ futhi kuthuthukiswe kakhulu ama-gross moves ngehora.
Le phrojekthi ihambisana nesu lesikhungo sokuguqula izinto zibe zedijithali, ukusebenza kahle, kanye nesibopho semvelo, ibonisa ukuthi izinhlelo zokubona zingasekela kanjani imisebenzi yezindaweni zemikhumbi ezisimeme.
Izinzuzo Ezingalinganiswa: Ukusebenza Kahle, Ukuphepha, Nokonga Izindleko
Ukusetshenziswa kwezinhlelo zokubona ezindaweni ezizenzakalelayo zeziteshi zemikhumbi yezitsha kuletha izinzuzo ezintathu eziyinhloko ezithinta ngqo imali yesiteshi:
1. Ukuthuthukiswa Kokusebenza Kahle
Izinhlelo zokubona ziqeda imingcele yesivinini sokusebenza ngesandla. Ngokunciphisa izikhathi zokujikeleza amaloli, ukukhulisa izinyathelo zekhreyini ngehora, nokwenza okuzenzakalelayo ukufakwa kwedatha, iziteshi zingacubungula iziqukathi eziningi ngengqalasizinda efanayo. Ngokwesibonelo, uhlelo lukaVisy eGothenburg luthuthukise ukudlula kwekhreyini, kanti isixazululo sikaHuaming Vision sinciphise isikhathi sokucubungula isiqukathi esisodwa ngo-92% (kusuka kumaminithi angu-3 kuye kumasekhondi angu-15). Lokhu kuthuthukiswa kokusebenza kubalulekile ekubhekaneni nosayizi omkhulu wokukhula wezikebhe zeziqukathi nasekufezeni izikhathi eziqinile zokuhlinzekwa.
2. Ukuphepha Okuthuthukisiwe kanye Nokuncishiswa Kwezingozi
Ukuphepha kuyinto ebaluleke kakhulu ekusebenzeni eziteshini, lapho imishini esindayo nabasebenzi abantu basebenza eduze kakhulu. Izinhlelo zokubona zixazulula lokhu ngokudala izindawo zokuphepha zikagesi ezibonakalayo ngokunemba okulinganiselwe ku-centimeter, ziqala izexwayiso ezizwakalayo nezibonakalayo uma abasebenzi bangena ezindaweni eziyingozi. Ziphinde zivimbele izingozi ngokugada ngesikhathi sangempela izindawo zezindiza, izitsha, namaloli.
Kwelodwa icala, imbobo yabika ukuncipha okungu-42% ezingozini zokushayisana kwezinto zokusebenza ngemuva kokusebenzisa uhlelo lokubona abantu nezimoto olusekelwe ku-AI kuma-reach stacker nama-forklift. Ukuthwebula kwezithombe okungu-360-degree kwalolu hlelo kususa izindawo ezingabonakali, kunikeze abasebenzisi isexwayiso sangaphambili ngezingozi ezingenzeka.
3. Izindleko Zokusebenza Eziphansi
Ngokuzenzakalela imisebenzi eyenziwa ngesandla, izinhlelo zokubona zinciphisa isidingo sabashayeli bamaloli abanekhono eliphezulu nabahlanganisi bamarekhodi. Ukuqeqesha umshayeli olulodwa olunekhono kungathatha iminyaka, futhi abaqhubi bamachweba omhlaba wonke babhekene nokushoda kwethalente elinjalo. Izinhlelo zokubona nazo zinciphisa izindleko zokugcinwa kwemishini ngokunciphisa ukungqubuzana nokuguga okubangelwa ukungahlangani kahle. Ngaphezu kwalokho, ikhono lazo lokuhlanganiswa nezinhlelo ezikhona zamachweba (TOS, PLC) kusho ukuthi amachweba akudingi ukufaka esikhundleni imikhumbi yamaloli ephelele, kunciphisa izindleko zokutshala imali zokuqala.
Amathrendi Esikhathi Esizayo: Isizukulwane Esilandelayo Sohlelo Lokubona Amakhenane
Njengoba ubuchwepheshe buqhubeka bukhula, izinhlelo zokubona zamakhenane ezizenzakalelayo ezimileyo zizoba nobuchwepheshe obuningi futhi zihlanganiswe kakhulu. Nansi imikhuba eyinhloko okufanele uyibheke:
1. Ukuhlanganiswa Kwezinsiza Eziningi kanye Ne-IoT
Izinhlelo zesikhathi esizayo zizohlanganisa idatha yokubona nezinputi ezivela kumasensori e-IoT, ukuxhumana kwe-5G, kanye ne-radar ukuze kudalwe "i-digital twin" ephelele yephothi. Lokhu kuzovumela ama-crane ukuthi alindele izinguquko ngesikhathi sangempela—njengokunyakaza kwezikhwama okubangelwa umoya noma izinguquko ezisheshayo endaweni ye-truck—futhi alungise imisebenzi ngaphambi kokuba zenzeke.
2. I-AI Ethuthukisiwe Yokugcinwa Okubikezelayo
Ngaphezu kokusebenza, izinhlelo zokubona zizobuye zisetshenziselwe ukugcinwa okubikezelayo. Ngokuhlaziya izithombe zezinto ze-crane (isibonelo, izintambo, izisabalalisi) ngezimpawu zokuguga noma umonakalo, amamodeli e-AI angaxwayisa amaqembu okugcinwa ngaphambi kokuba kube nezinkinga, kunciphise isikhathi sokungasebenzi esingahlosiwe.
3. Ukuzivumelanisa Okukhulu Nezimo Eziyingozi
Ukuqhubeka kwezinguquko ezithuthukisiwe kumalulwazi okucubungula izithombe kuzothuthukisa ukusebenza kwesistimu ezimeni zezulu ezinzima, njengenkungu enkulu, imvula, neqhwa. Izinhlelo zesizukulwane esilandelayo zizosebenzisa ukukhanyisa okuzivumelanisa nezimo kanye namalulwazi athuthukisiwe okususa inkungu/imvula ukuze kugcinwe ukunemba okuphezulu kungakhathalekile izinto ezithinta imvelo.
4. Ukuhlanganiswa Okungenamathungo Nezinhlelo Zokuhlinzekwa Kwempahla Emhlabeni Wonke
Izinhlelo zokubona zizoba nendima ebalulekile ekwandiseni ukuveza kahle kwezimali zokuhamba. Ngokuthwebula nokudlulisa idatha ye-container (ID, isimo, indawo) ngokuzenzakalelayo ezinkundleni zokuhamba zomhlaba, zizovumela ukulandela isikhathi sangempela kwezimpahla ukusuka emkhunjini ziye endaweni, kunciphisa ukubambezeleka nokuthuthukisa ukuphathwa kwezimpahla.
Isiphetho: Ukutshala Izinhlelo Zokubona Ukuze Ube Nokuncintisana Kwesikhathi Eside
Izinhlelo zokubona azisewona into ewubukhazikhazi ezindaweni zemikhumbi ezizenzakalelayo—zibalulekile ukuze uhlale unqobayo emhlabeni wokuhweba womhlaba wonke osheshayo. Ngokuxazulula ukungasebenzi kahle kokusebenza ngesandla, ukuthuthukisa ukuphepha, nokunciphisa izindleko, lezi zinhlelo ziguqula imisebenzi yezindaweni zemikhumbi futhi zinika amandla isizukulwane esilandelayo sezindawo ezihlakaniphile, ezisimeme.
Kubasebenzisi bezindaweni zemikhumbi abacabanga ngokuzenzakalela, ukutshala imali ezinhlelweni zokubona ezithuthukisiwe kufanele kube yinto ehamba phambili. Izinzuzo ezibonakalayo—izikhathi ezisheshayo zokujikeleza, ukukhiqiza okuphezulu, kanye nezindleko zokusebenza eziphansi—zidlula kakhulu ukutshalwa kwezimali kwasekuqaleni. Njengoba ubuchwepheshe buqhubeka nokuthuthuka, izinhlelo zokubona zizoba yingxenye ebaluleke kakhulu emisebenzini yezindaweni zemikhumbi, zixhume ama-crane kulo lonke uchungechunge lokuhlinzekwa futhi ziqhube ikusasa lezinto zokuhambisa izimpahla zomhlaba wonke.
Noma ngabe ubuyekeza isiteshi esikhona noma wakha isiteshi esizenzakalelayo esisha, uhlelo olufanele lokubona lungavula amazinga angakaze abekho ngaphambili okusebenza kahle nokwethembeka. Ikusasa lokusebenza eziteshini zemikhumbi yezitsha libonakala—futhi selifikile.
Uxhumane
Sicela uxhumane nathi uhambele
WhatsApp
WeChat