isiZulu
Amamojula weKhamera kuma-AR/VR Izinsiza: Iziqondiso Eziyinhloko Ezihlela Ikusasa Lezobuchwepheshe Obujulile
Kwadalwa ngo 10.28
I-AR/VR imboni ibhekene nokukhula okungakaze kubonwe, i-Statista ibikezela ukuthi usayizi wemakethe yomhlaba uzofinyelela ku-$48.8 billion ngonyaka ka-2026. Enhliziyweni yokwandiswa kwalokhu kukhona ingxenye ebalulekile evame ukungabhekwa ngabasebenzisi bokugcina:amamojula ekhanda. Lezi zinhlelo ezincane kodwa ezinamandla ziyiz "amehlo" wezinsiza ze-AR/VR, zenza konke kusukela ekuqapheleni ukuhamba kuya ekuhloleni imvelo. Njengoba ubuchwepheshe bokuzijabulisa buqhubeka, amamojula wekhamera ayashintsha ngokushesha ukuze ahlangabezane nezidingo zokuhlangenwe nakho kwe-AR/VR okungcono, okuphendulayo, nokufinyeleleka. Kulesi sihloko, sihlola ukuthrendwa okukhulu okwamanje kumamojula wekhamera ye-AR/VR kanye nemiphumela yawo embonini.
1. Ukunciphisa Usayizi Ngaphandle Kokwehlisa Ukusebenza
Omunye wezinkinga eziphuthumayo kubakhiqizi bezinsiza ze-AR/VR ukuhlela ubukhulu bokwakheka nokusebenza. Ikhanda le-AR lokuqala, ikakhulukazi, lalikhulu futhi lingakhululekile, ikakhulu ngenxa yezingxenye zekhanda ezinkulu. Namuhla, umkhuba ubonakala ngokucacile ekunciphiseni, okuholelwa yizidingo zabathengi zokuthi kube nezinsiza ezilula, eziphathekayo ezingasetshenziswa amahora ngaphandle kokukhathala.
Abakhiqizi abahamba phambili bezengezo bafeza lokhu ngokusebenzisa ubuchwepheshe obuphambili bokwakha obuncane. Isibonelo, amamojula amasha e-AR kaQualcomm anemikhawulo ye-5x5mm kuphela, okunciphisa ngosizo lwe-40% uma kuqhathaniswa namamodeli avela ku-2022. Nokho, le mncintiswano ayiphumi engozini yokusebenza. Lezi zinsiza ezincane zisagcina izinga eliphezulu lokuhamba (kuze kube yi-120fps) kanye nezibuko ezibanzi (FoV)—okubalulekile ukuze kuthathwe umqondo ophelele wendawo yomsebenzisi.
Umthelela wale mikhuba ubonakala emikhiqizweni yokusetshenziswa. I-Quest 3 headset ye-Meta, ekhishwe ngo-2023, ifaka imodyuli ezine zekhamera ezincane ezincane ezingama-30% zincane kunezika-Quest 2, kodwa zikhombisa ikhwalithi yokudlula engcono. Le miniaturization iphinde yavula amathuba ezigqoko ze-AR, njenge-XREAL Air 2, ezinezinhlelo ezinhle ezifanayo nezibuko zelanga ezijwayelekile, ngenxa enkulu yemodyuli zekhamera ezincane, ezisebenza kahle.
2. Ukuphakama Kwezixazululo Eziphakeme Nokuhamba Okusheshayo
Njengoba okuqukethwe kwe-AR/VR kuba nekhono elikhulu, abasebenzisi balindele izithombe ezifanayo nezempilo yangempela—futhi amamojula wekhamera ayavuma le nselelo ngokukhuphula isixazululo kanye ne-dynamic range. Amakhamera ajwayelekile e-AR/VR abekwe phezulu ku-1080p resolution, kodwa amamojula e-4K manje asejwayelekile kumadivayisi aphakathi nendawo kuya phezulu, kanti okukhethwa ku-8K kuyavela emishini ye-professional-grade.
I-High resolution iyashintsha emisebenzini ebalulekile ye-AR/VR. Isibonelo, ama-4K camera modules kumadivayisi e-medical AR avumela odokotela ukuba babone izithombe ezinemininingwane ze-anatomy ezibekwe phezu komzimba wesiguli ngokucacile okungakaze kubonwe. Ku-VR, i-high resolution passthrough (ikhono lokuthi "ubone ngaphezulu" kwe-headset emhlabeni wangempela) ikhipha "umphumela wesikrini"—umfanekiso ongenamsebenzi obonakalayo osanda kwaba yinkinga kumadivayisi aphilayo.
I-Dynamic range iyindawo ethuthukile. Ama-modules we-AR/VR amanje angabhekana nezinguquko ezinkulu ekukhanyeni, kusukela ekukhanyeni kwelanga ngaphandle kuya ezindaweni ezimnyama ngaphakathi, ngaphandle kokweqisa noma ukunciphisa ukukhanya kwemifanekiso. Lokhu kubalulekile ezinhlelweni ze-mixed reality (MR), lapho izinto ezibonakalayo kufanele zihlangane kahle nezwe langempela. Izinkampani ezifana ne-Sony zihola phambili lapha, ngama-sensors akamuva e-IMX890 anikeza ama-stop angu-14 we-dynamic range, ukuthuthukiswa okungu-27% uma kuqhathaniswa nezizukulwane ezedlule.
3. Ukuhlanganiswa Kwezinsiza Eziningi Ukuze Kwandiswe Ukuqaphela Kwemvelo
Sekuphelile izinsuku zokusebenzisa ama-AR/VR anekhamera eyodwa. Izinsiza zanamuhla zisebenzisa ukuhlanganiswa kwezinsiza eziningi—zihlanganisa idatha evela kumakhamera amaningi, kanye nezinye izinsiza ezifana nama-accelerometers nama-gyroscopes—ukwakha ukuqonda okuphelele kwemvelo yomsebenzisi. Le mikhuba iphushwa yidingo lokulandela kahle, ukuqashelwa kwezinto okungcono, nokungena okungcono.
I-AR/VR headset ejwayelekile ephezulu manje ifaka inhlanganisela yezinhlobo zamakhamera: amakhamera e-RGB ukuze kubonwe imibala, amakhamera okuphakama ukuze kukalwa ubude, namakhamera e-infrared (IR) ukuze kulandelwe ezimeni zokukhanya okuphansi. Isibonelo, i-Apple Vision Pro isebenzisa ama-module amakhamera angama-12, kufaka phakathi amakhamera amabili e-6MP RGB, amakhamera amane okuphakama, namakhamera amathathu e-IR, ukuze iqhube izici zayo zokubala indawo. Ngokuhlanganisa idatha evela kulezi zinsiza, i-headset ingalandela ukuhamba kwamehlo komsebenzisi, izimpawu zezandla, kanye nesikhundla somzimba ngokuqonda okuncane kakhulu.
I-fusion ye-multi-sensor iphinde ivumele ukuhlela kwemvelo okuqhubekayo. I-SLAM (Simultaneous Localization and Mapping), ubuchwepheshe obuvumela amadivayisi ukuthi ahlele izindawo ezingaziwa ngenkathi elandelela indawo yawo, ithuthukiswa kakhulu ngamakhamera amaningi. Ngemininingwane evela ezindaweni ezahlukene, ama-algorithms e-SLAM angakha amapulani e-3D anembile, anemininingwane eminingi, okubalulekile ezinhlelweni ezifana nokwakhiwa kwangaphakathi kwe-virtual kanye nokuqeqeshwa kwe-AR kwezimboni.
4. Ukuklama Okuphansi Kwamandla Ukuze Kwandiswe Impilo Yebhethri
Impilo yebhethri ibiyinkinga njalo kumadivayisi e-AR/VR. Imodyuli yekhamera ingezinye zezingxenye ezidla amandla kakhulu, njengoba zihlala ziqopha futhi zisebenza ngedatha. Ukuze kubhekwane nalokhu, abakhiqizi bagxila kumiklamo enezindleko eziphansi kumamodyuli abo amasha ekhamera—umkhuba ophawulekayo njengoba amadivayisi e-AR/VR eqhubeka nokuba lula.
Izobuchwepheshe eziningi zenza le shintsho. Omunye wabo ngu-pixel binning, ophatha idatha evela kumapixels amaningi ukuze unciphise inani lokucubungula elidingekayo, ngaleyo ndlela kwehlise ukusetshenziswa kwamandla. Isibonelo, ama-sensors e-OmniVision’s OV6211 asebenzisa i-4-in-1 pixel binning ukuze anikeze isixazululo se-1080p ngama-50mW wamandla, okungama-50% wamandla uma kuqhathaniswa nezinketho ezingekho ku-binned.
Enye inovatheki iyizinga lokushintsha elihambisanayo. Amamojula wekhamera manje angakwazi ukulungisa izinga lawo lokushintsha ngokuya ngomsebenzi okukhona—sebenzisa i-120fps yemidlalo ye-VR esheshayo bese wehla ku-30fps kwezicelo ze-AR ezimile njengefunda umbhalo. Lokhu kushintsha okujwayelekile kunganciphisa ukusetshenziswa kwamandla ngaphezu kuka-35%, ngokusho kokuhlolwa kwe-VR/AR Association.
Izinzuzo zamamojula amakhamera aphansi amandla zicacile. Abasebenzisi bezibuko ze-AR zakamuva manje bangajabulela kuze kube yimizuzu engu-6 yokusebenzisa okuqhubekayo ngokuqhuba okukodwa, phezulu kokuphuma kwemizuzu engu-2-3 eminyakeni emibili edlule. Kubakaki be-VR, ukuphila kwesibambo eside kusho ukuncipha kokuphazamiseka phakathi nezikhathi zokudlala ezinde noma emihlanganweni yokusebenza.
5. Ukuhlanganiswa kwe-AI ukuze Kwenziwe UkuProcessing Okuhlakaniphile
Ubuhlakani bokwenziwa (AI) bujolisa emkhakheni wezobuchwepheshe, futhi ama-module wekhamera ye-AR/VR awahlukanisiwe. Ama-module wamanje aqhubeka nokufaka ama-chips e-AI kumadivayisi ukuze avumele ukucubungula okuhlakaniphile ngesikhathi sangempela, kunciphisa ukuthembela ekubambeni kwedatha efwini futhi kuthuthukise isikhathi sokuphendula.
Amamojula ekhanda anesizindalwazi se-AI akhombisa kahle ekuqondeni izinto nasekuqondeni izimo. Isibonelo, uhlelo lokusebenza lwe-AR lokuthenga lungasebenzisa imojula yekhamera enesizindalwazi se-AI ukuze ikwazi ngokushesha ukuhlonza umkhiqizo umsebenzisi ophakathi kokuwuphethe futhi ibonise ulwazi oluhambisanayo (njengokuhlanganiswa kwamanani noma ukubuyekezwa) okukhombisa phezu kwento. Ezindaweni zezimboni, amamojula ekhanda anesizindalwazi se-AI angakwazi ukuthola amaphutha emishini ngesikhathi sokuhlola ukugcinwa okuqondiswe yi-AR, ebika kubachwepheshe ngezinkinga ngaphambi kokuba zikhule.
AI futhi ithuthukisa ukuxhumana komsebenzisi. Ama-modules ekhamera anama-AI angakwazi ukuqaphela izimpawu zokuhlanganyela nezimo zobuso ngaphandle kwesidingo sokulawulwa okwengeziwe. Ukulandela izandla kwe-Meta Quest 3, okuqhutshwa ngedatha yekhamera eyenziwe ngama-AI, kuvumela abasebenzisi ukuthi baphathe izinto ezibonakalayo ngezinyathelo zemvelo, ezifana nokucindezela nokudonsa. Lezi zinga lokuqonda bekungakwenziwanga ngezinhlelo zekhamera zangaphambili, ezingekho ku-AI.
I-AI esebenzisa idivayisi iphinde ibhekane nezinkinga zokuphepha. Ngokucubungula idatha endaweni (esikhundleni sokuyithumela efwini), amamojula wekhamera anciphisa ingozi yokuthi ulwazi olubucayi lungavuleka. Lokhu kuyisici esibalulekile sokuthengisa izixazululo ze-AR/VR zamabhizinisi, lapho ukuphepha kwedatha kuyinto ebalulekile kakhulu.
6. Intuthuko kuTekhnoloji Yokuzwa kwe-3D
3D sensing ibalulekile ekwakheni okuhlangenwe nakho kwe-AR/VR okuyiqiniso, njengoba ivumela amadivayisi ukuba abone ukujula nokwakheka komhlaba wangempela. Eminyakeni yamuva, kube nokuthuthuka okukhulu kumamojula wekhamera ye-3D sensing, nezobuchwepheshe ezimbili ezihola phambili: ukukhanya okuhlelwe kahle kanye nesikhathi sokuhamba (ToF).
Izinhlelo zokukhanya ezakhiwe zikhombisa iphethini yamadoti noma imigqa endaweni ethile futhi zisebenzisa ikhamera ukuze ziqophe ukuthi iphethini ibhujiswe kanjani. Le bhujiswa isetshenziswa ukuze kubalwe ukujula. I-Apple isesebenzisa ukukhanya okuhlelekile kuhlelo lwayo lwe-Face ID isikhathi eside, futhi le teknoloji manje isiqala ukufika kumadivayisi e-AR/VR. Ukukhanya okuhlelekile kunikeza ukunemba okuphezulu (kuze kube ngu-1mm) kodwa kunomkhawulo ngezinga, ngokuvamile kusebenza kangcono phakathi kuka-2 meters.
Ubuchwepheshe be-ToF, ngokuphambene, b measuring isikhathi esithathwa ukukhanya ukuhamba ukusuka kukhamera ukuya endaweni bese buhamba buya. Lokhu kuvumela ukuhamba kwesikhathi eside sokuhlola i-3D (kuze kube yimitha eyi-10) futhi kusebenza kahle ezimeni zokukhanya ezahlukene. Amamojula amasha e-AR kaSamsung asebenzisa ubuchwepheshe be-ToF ukuze avumele ukuhlela kahle kwezindawo, okwenza kube kuhle ezindaweni ezinkulu ze-VR ezifana nemicimbi ye-virtual.
Ukwakhiwa okusha ku-3D sensing kuyi-LiDAR (Light Detection and Ranging) integration. Ama-sensor e-LiDAR, asebenzisa ama-pulse e-laser ukukala ibanga, ahlanganiswa namakhamera ajwayelekile ukuze akhe imephu ye-3D enemininingwane eminingi. I-module yekhamera ye-Apple Vision Pro esebenzisa i-LiDAR ingakwazi ukuhlela igumbi ku-3D ngaphansi kwemizuzwana emibili, ivumela izinto ezivirtual ukuba zisebenze nezindawo zangempela (njengokuthi ikomishi evirtual ibekwe etafuleni langempela) ngokweqiniso okukhulu.
Isiphetho: Ikusasa lezi zinhlelo ze-AR/VR zeKhamera
Iziqhamo ezihlanganisa ama-module wekhamera ye-AR/VR—ukunciphisa, ukwandisa isixazululo, ukuhlanganiswa kwezinsiza eziningi, imiklamo enezindleko eziphansi, ukuhlanganiswa kwe-AI, kanye nokuhlola okuqhubekayo kwe-3D—konke kuhamba phambili ngenhloso evamile: ukudala okuhlangenwe nakho kwe-AR/VR okungcono, okucacile, nokufinyeleleka kalula. Njengoba lezi zobuchwepheshe ziqhubeka nokuthuthuka, singalindela ukubona izinhlelo eziningi eziphawulekayo, kusukela kwezempilo nasemfundweni kuya kokuzijabulisa nasezinkampanini.
Kubathengi, lokhu kusho amadivayisi alula, anokuthokozisa okuningi nezithombe ezihambisana nempilo yangempela. Kubhizinisi, kusho amathuluzi anamandla kakhulu okuqeqesha, ukuklama, nokubandakanya amakhasimende. Futhi embonini ye-AR/VR njengokuphelele, imodyuli yekhamera izohlala iyinsiza ebalulekile yokwakha, iqhuba imingcele yalokho ubuchwepheshe obujulile bungakufeza.
Njengoba sibheka phambili, into eyodwa icacile: "amehlo" ezinsizakusebenza ze-AR/VR ayanda, aqonda kahle, futhi asebenza kahle—futhi lokho kuwumsebenzi omuhle kunoma ubani ofuna ukungena emhlabeni ophilayo.
Uxhumane
Sicela uxhumane nathi uhambele
WhatsApp
WeChat