Yayin da fasahar robotics ke ƙaruwa daga sarrafa kansa zuwa ga hankali na gabaɗaya, wani sabon nau'i da aka sani da "robot ɗin EGO" yana sake fasalta yadda injuna ke hulɗa da muhallinsu. Ba kamar shirye-shiryen da ba na intanet ba na makamai na masana'antu na gargajiya ko kuma fahimtar robots masu aiki ba, robot ɗin EGO (egocentric) yana ɗaukar hangen nesa na mutum na farko a matsayin asalin fahimta, yana cimma microsecond, daidaitaccen hannu, daidaitawar ido da kuma lura mai aiki.Babban kayan aikin da ke ba da damar wannan juyin juya halin-kyamarar hangen nesayana canzawa daga "ido na lantarki" mai tsayayye zuwa kansa na robot-hanyar sadarwa ta wayar da kan jama'a. Babban haɗin kai inganci module ɗin kyamarar sitiriyo ba wai kawai na'urar ɗaukar hoto ba ce; ita ce ginshiƙin gina ainihin 3D module na kyamara don robot.
1. Nau'ikan Robot guda uku, Hanyoyi uku na "Ganin Duniya"
Robot na gargajiya (misali, sashin masana'antu, AGV): Ba a haɗa shi da intanet ba-Mai aiwatarwa wanda aka tsara wanda ke maimaita ayyuka ba tare da an gani ba. Kyamarar tana da ƙarfi kuma tana fitar da sigina masu sauƙi kawai, robot ɗin ba zai iya daidaita yanayin tafiyarsa a ainihin lokacin ba.Robot mai jiki (misali, robot mai hidima, mai ruɓi huɗu): An ɗora kyamarar a jikin robot ɗin, rufe fahimta-madaurin aiki. Duk da haka, har yanzu akwai jinkiri na "gani"→lissafi→motsa" kuma babu hangen nesa mai aiki. Kyamarar robot mai tsari ta fara buƙatar zurfin bayani da kuma saurin firam mai girma.
Robot na EGO (robot mai tsayin daka): Kyamarar'ra'ayinsa shine na robot kai-asali. Shi yana canza yanayin kansa ko jikinsa sosai don inganta yanayinsa sokarkatarwa, matsowa kusa, juyawa a gefe. Kowane firam ɗin hoto ya yi daidai da kusurwoyin haɗin gwiwa, saurin da kuma kwararar ruwa a wannan lokacin, suna samar da cikakken kai-ƙwaƙwalwar ajiya. Dole ne na'urar kyamarar robot ta ego ta musamman ta cika ƙa'idodin kayan aiki masu tsauri don daidaitawa da kewayon motsi.
2. Tsalle-tsalle Uku a Matsayin Kyamarar: Daga Kayan Aiki zuwa Fuskar Sadarwa
Mataki na 1 (robot na gargajiya): Mai lura da kyau-ƙarancin buƙatun fasaha (monocular, birgima rufewa, ƙarancin kewayon motsi).
Mataki na 2 (robot mai jiki): Sashe na jiki- yana buƙatar cikakken bayani, 30-60 fps kuma ya fara buƙatar rufewa ta duniya da HDR.
Mataki na 3 (robot ɗin EGO): Kai-hanyar sadarwa ta wayar da kan jama'a dole ne a gamsar da shi: microsecond matakinDaidaita sitiriyo, color rufewa ta duniya, HDR > 90dB, matsananci-ƙarancin karkacewa, faɗi-Gilashin kusurwa da kuma daidaita sitiriyo na masana'anta. Waɗannan fasalulluka su ne ainihin abin da cikakken tsarin kyamarar sitiriyo da tsarin kyamarar rufewa na duniya ya kamata su bayar.
3. Muhimman Siffofin Tsarin Binocular da aka ƙera don Robots na EGO
Wannan na'urar kyamarar binocular samfurin 3290 kuma manufa ce ta 3D Module ɗin kyamara don na'urar hangen nesa ta robot da robotics. Manyan fasalulluka sun haɗa da:
Biyu 1/2.6''kololuwar rufewar duniya CMOS: Duk pixels da aka fallasa a lokaci guda, suna kawar da birgima murƙushewar rufewa yayin motsi kai mai sauri; manyan pixels (3.0)μm) tabbatar da ƙarancin inganci-haske SNR-ainihin tsarin kyamarar rufewa ta duniya.
3840×1080P @ 30 fps:Tashoshi masu zaman kansu na 1080P a kowane ido, kayan aiki Daidaita firam ɗin matakin (<50μs), yana ba da damar gaske-Kimanta zurfin lokaci. Wannan ƙudurin ya sa na'urar ta zama kyamarar hangen nesa ta robot ta gaske.
91.56dB HDR (layi mai layi 61.56dB):Yana riƙe da cikakkun bayanai game da haske da inuwa a ƙarƙashin canje-canjen haske mai tsanani (100) kayan more rayuwa na cikin gida zuwa 100,000 hasken rana mai haske), yana hana algorithm ɗin kasancewa makanta.
Biyu 125°matuƙar faɗi, kusurwa matuƙar, ƙarancin karkacewa: Ya ƙunshi faffadan fage na gani daga bene zuwa sama da kai tare da karkatar da talabijin <3%. Wannan yana inganta ƙuntatawa na epipolar kuma yana ba da ƙwarewar sitiriyo mai daɗi don aikin VR ta hanyar sadarwa, yana biyan buƙatun gani na na'urar kyamarar robot ta ego gaba ɗaya.
Daidaita sitiriyo na masana'anta: Kowace na'ura tana ba da sigogi na ciki, ma'aunin murdiya, ma'aunin gyaran sitiriyo da matrices na hasashen. Masu amfani za su iya yin triangulation kai tsaye, ƙirƙirar taswirar zurfi da sake gina gajimare mai maki 3D-ainihin ƙarig and wasa na 3D module na kyamara don robot.
Filogin USB 2.0 kuma yi wasa + Karfin jituwa da Windows/Linux: Direban UVC-kyauta, cikakken SDK da kuma maɓallan ROS/ROS2. Tare da tallafin IMU na waje da direbobin ROS2 na asali, shi ma wani tsarin kyamara ne da ake tallafawa ta Imu & ros2 don hangen nesa na robot., ai mai hadewa, tsarin slam, a shirye don gani odometry na inertial (VIO) da ainihin taswirar lokaci.
4. Yanayin Aikace-aikace na yau da kullun
Koyon Kwaikwayo don robots na EGO: Sanya module ɗin (a matsayin module na kyamarar robot na ego) a kan wuyan hannu ko kai na robot maƙallin da aka sata don tattarawa da farko Bidiyon mutum na ayyukan ɗan adam. Bidiyon yana ɗauke da cikakken zurfin lokaci da tambarin lokaci, wanda ke ba da damar algorithms na koyon manufofi su fitar da umarni na haɗin gwiwa kai tsaye.
Sadarwar VR da kuma tagwayen dijital: Ma'aunin azaman tsarin kyamarar sitiriyo da tsarin kyamarar 3D don robot Yana fitar da kwararar bidiyo na hagu/dama zuwa na'urar kai ta VR tare da ƙarancin latency. Ana mayar da motsin kai ga robot ɗin, yana ba da damar duba "daga nesa"-a kusa da "ikon sarrafawa yana da mahimmanci ga duba mai zurfi ko ma hanyoyin likita.
Babban sarrafa damar shiga tsaro da gano rayuwa: Yana amfani da bayanai masu zurfi don bambance fuskoki na gaske daga hare-haren hoto/bidiyo kuma yana fitar da samfurin fuska na 3D don ƙarin tabbatarwa. A nan na'urar tana aiki azaman kyamarar hangen nesa ta robot.
Fahimtar zurfin fahimta da kuma guje wa cikas ga robots ɗin sabis: A otal-otal, asibitoci, gidajen cin abinci, tsarin yana aiki a matsayin babban firikwensin gani (module na kyamarar hangen nesa ta robotics), yana samar da ingantaccen zurfin daga 0.2 m zuwa 5 m don VSLAM da ainihin Gujewa cikas na lokaci. 125°FOV mai faɗi yana rage wuraren makafi kuma HDR yana tabbatar da aiki mai kyau kusa da bangon labulen gilashi ko hasken rana kai tsaye. Wannan yanayin kuma yana amfana daga matsewar haɗin da aka bayar ta hanyar na'urar kyamara mai goyan bayan imu & ros2 don hangen nesa na robot., ai mai hadewa, tsarin slam.
5. Hasashen Nan Gaba: Daga "Ido" Zuwa "Kwakwalwa"-Haɗa Ido”
Tare da ci gaba a cikin ƙididdigar gefe da kwakwalwan neuromorphic, robots na EGO na gaba za su yi aikin sarrafa gani kai tsaye kusa da firikwensin, suna cimma microsecond matakin "pixel" to "juyawa" action. Tsarin binocular ɗinmu ya riga ya samar da shigarwar abubuwan jan kayan aiki da kuma abubuwa da yawa hanyoyin haɗa kyamara, a shirye don dandamalin kwamfuta daban-daban (FPGA, GPU, masu haɓaka AI). An fara tattarawa da yawa Bidiyon hangen nesa na mutum daga robot na EGO, waɗanda aka yi musu bayani tare da umarnin aiki masu dacewa, zasu taimaka wajen horar da hangen nesa gabaɗaya samfuran tushen aiki yayi kama da manyan samfuran harshe don rubutu. kyamarori masu inganci na binocular sune mataki na farko a cikin wannan bayanai juyin juya hali da aka yi.
Bawa robot ɗinka farko idanun mutum fara da na'urar hangen nesa ta mu.
Idan kana son ƙarin bayani, tuntuɓe mu.
Lokacin Saƙo: Mayu-29-2026
