Imodeli ejwayelekile ye-unithi yokucubungula isithombe se-thermal imaging camera

Ihadi ye-thermal imaging.kamerais key to converting digital signal, which has been signal processed, into an intuitive thermal image. Common models perform exceptionally well in various fields.

ingka Technology's LC221 ke modulu kamera infrared e phahlo. E sebelisa tharollo e advanced ASIC, e kopaneng le mocheso o sa futhumetseng o telele infrared vanadium detector, e fa mekhahlelo e phahameng ea 384×288 kapa 640×512. Unit ea eona e matla ea ts'ebetso ea setšoantšo e tšehetsa NUC correction ho felisa ho se lekane ha khanya ea setšoantšo, hammoho le mekhahlelo e fapaneng joalo ka ho tlosoa ha li-pixel tse mpe le ho fokotsa molumo oa nako le sebaka ho ntlafatsa boleng ba setšoantšo. Pseudo-color processing e ka bontša ka mokhoa o hlakileng karohano ea mocheso. Hape, e tšehetsa H264/H265 video encoding, e etsa hore ho be bonolo ho boloka le ho fetisa. E sebelisoa hangata maemong a joalo ka ho batla libaka tse chesang le ho lebisa mocheso, joalo ka ho sibolla ba kotsi ho laola ts'ireletso le ho lemoha ho senyeha ha lisebelisoa tsa indasteri.

FLIR's imahlanganisi zezimveliso zifaka ama-chips aphezulu wokucubungula izimpawu zedijithali, anikela ngejubane elisheshayo lokucubungula. Ngokwandisa ibanga le-grayscale elihambelana ne-histogram, umehluko omncane wezinga lokushisa ubonakala. Imodeli yokulungisa engaphelele enembile iqinisekisa ukukala kwezinga lokushisa okunembile nokukhanya okufanayo kanye nemibala endaweni efanayo yezinga lokushisa lesithombe. Iphinde isekele ukuhlanganiswa, ihlanganisa imifanekiso ye-thermal nemifanekiso yokukhanya okubonakalayo, okusetshenziswa kabanzi emikhakheni edinga ukunemba okuphezulu ekuhlaziyeni izithombe, njengokuhlola imboni nokuphathelene namandla, ukuze kuhlolwe kahle isimo sezinsiza.

Izi zikhangiso zokucubungula isithombe zisebenzisa ubuchwepheshe bokubala obuphindaphindiwe kanye nezinhlelo ezithuthukisiwe, amandla aqinile okucubungula ngesikhathi sangempela nokuphothula inani elikhulu lokucubungula idatha yokushisa esikhathini esifushane. Ngokuhlanganisa izinhlelo zokufunda ezijulile nokusebenzisa amanethiwekhi e-neural convolutional, ingakwazi ukuzithola ngokuzenzakalelayo izinto ezihlosiwe nokuhlola ukungajwayelekile kokushisa ngokusekelwe ezicini zokushisa. Ekubhekeni kwezinkanyezi zomlilo, ingakwazi ukusiza ngokushesha ukuthola imithombo yomlilo kanye nomusi; ekuvikelekeni, ingahlaziya ukuziphatha okungajwayelekile kanye nezinga lokushisa kwabantu. Idlala indima ebalulekile ezimweni zokubhekwa ezinkulu, ezinde ezifana nokubhekwa kokuphepha kanye nomlilo wehlathi, iqinisekisa ukuphepha kwendawo.

Lezi zindlela ezivamile zezingxenye zokucubungula izithombe, ezinezinzuzo zazo zobuchwepheshe, zikhuthaza ukusetyenziswa nokuthuthukiswa kobuchwepheshe bokuthwebula izithombe ezishisayo emikhakheni efana nezokuphepha. Esikhathini esizayo, ngokuthuthuka kobuchwepheshe, kuzovela amamodeli aphezulu kakhulu, okuzokwandisa kakhulu ukusebenza nokubaluleka kokusetshenziswa kwamakhamera okuthwebula izithombe ezishisayo.