{"id":11549,"date":"2025-05-04T21:00:25","date_gmt":"2025-05-04T19:00:25","guid":{"rendered":"https:\/\/geopard.tech\/?p=11549"},"modified":"2025-05-04T21:00:25","modified_gmt":"2025-05-04T19:00:25","slug":"cmtnet-omdefinierar-precisionsjordbruk-genom-att-overtraffa-traditionell-grodklassificering","status":"publish","type":"post","link":"https:\/\/geopard.tech\/swe\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/","title":{"rendered":"CMTNet omdefinierar precisionsjordbruk genom att \u00f6vertr\u00e4ffa traditionell gr\u00f6dklassificering"},"content":{"rendered":"

Noggrann klassificering av gr\u00f6dor \u00e4r avg\u00f6rande f\u00f6r modernt precisionsjordbruk, vilket g\u00f6r det m\u00f6jligt f\u00f6r jordbrukare att \u00f6vervaka gr\u00f6dors h\u00e4lsa, f\u00f6ruts\u00e4ga avkastning och f\u00f6rdela resurser effektivt. Traditionella metoder k\u00e4mpar dock ofta med komplexiteten i jordbruksmilj\u00f6er, d\u00e4r gr\u00f6dor varierar kraftigt i typ, tillv\u00e4xtstadier och spektrala signaturer.<\/p>\n

Vad \u00e4r hyperspektral avbildning och CMTNet Framework?<\/h2>\n

Hyperspektral avbildning (HSI), en teknik som samlar in data \u00f6ver hundratals smala, sammanh\u00e4ngande v\u00e5gl\u00e4ngdsband, har blivit banbrytande inom detta omr\u00e5de. Till skillnad fr\u00e5n vanliga RGB-kameror eller multispektrala sensorer, som samlar in data i ett f\u00e5tal breda band, ger HSI ett detaljerat "spektralfingeravtryck" f\u00f6r varje pixel.<\/p>\n

Till exempel reflekterar frisk vegetation starkt n\u00e4ra-infrar\u00f6tt ljus p\u00e5 grund av klorofyllaktivitet, medan stressade gr\u00f6dor uppvisar tydliga absorptionsm\u00f6nster. Genom att registrera dessa subtila variationer (fr\u00e5n 400 till 1 000 nanometer) med h\u00f6g rumslig uppl\u00f6sning (s\u00e5 fin som 0,043 meter) m\u00f6jligg\u00f6r HSI exakt differentiering av gr\u00f6darter, sjukdomsdetektering och jordm\u00e5nsanalys.<\/p>\n

Trots dessa f\u00f6rdelar st\u00e5r befintliga tekniker inf\u00f6r utmaningar n\u00e4r det g\u00e4ller att balansera lokala detaljer, som bladstruktur eller jordm\u00f6nster, med globala m\u00f6nster, s\u00e5som storskalig gr\u00f6df\u00f6rdelning. Denna begr\u00e4nsning blir s\u00e4rskilt tydlig i brusiga eller obalanserade datam\u00e4ngder, d\u00e4r subtila spektrala skillnader mellan gr\u00f6dor kan leda till felklassificeringar.<\/p>\n

F\u00f6r att hantera dessa utmaningar utvecklade forskare\u00a0CMTNet<\/strong>\u00a0(Convolutional Meets Transformer Network), ett nytt ramverk f\u00f6r djupinl\u00e4rning som kombinerar styrkorna hos faltningsneurala n\u00e4tverk (CNN) och Transformers. CNN \u00e4r en klass av neurala n\u00e4tverk utformade f\u00f6r att bearbeta rutn\u00e4tsliknande data, s\u00e5som bilder, med hj\u00e4lp av filterlager som uppt\u00e4cker rumsliga hierarkier (t.ex. kanter, texturer).<\/p>\n

\"CMTNet-arkitektur<\/p>\n

Transformatorer, ursprungligen utvecklade f\u00f6r naturlig spr\u00e5kbehandling, anv\u00e4nder sj\u00e4lvuppm\u00e4rksamhetmekanismer f\u00f6r att modellera l\u00e5ngsiktiga beroenden i data, vilket g\u00f6r dem skickliga p\u00e5 att f\u00e5nga globala m\u00f6nster. Till skillnad fr\u00e5n tidigare modeller som bearbetar lokala och globala funktioner sekventiellt anv\u00e4nder CMTNet en parallell arkitektur f\u00f6r att extrahera b\u00e5da typerna av information samtidigt.<\/p>\n

Denna metod har visat sig vara mycket effektiv och uppn\u00e5tt toppmodern noggrannhet p\u00e5 tre stora UAV-baserade HSI-dataset. Till exempel, p\u00e5 WHU-Hi-LongKou-datasetet, n\u00e5dde CMTNet en total noggrannhet (OA) p\u00e5 99,58%, vilket \u00f6vertr\u00e4ffade den tidigare b\u00e4sta modellen med 0,19%.<\/p>\n

Utmaningar med traditionell hyperspektral avbildning inom jordbruksklassificering<\/h2>\n

Tidiga metoder f\u00f6r att analysera hyperspektral data fokuserade ofta p\u00e5 antingen spektrala eller rumsliga egenskaper, vilket ledde till ofullst\u00e4ndiga resultat. Spektrala tekniker, s\u00e5som principal component analysis (PCA), minskade datas komplexitet genom att fokusera p\u00e5 v\u00e5gl\u00e4ngdsinformation men ignorerade rumsliga relationer mellan pixlar.<\/p>\n

PCA, till exempel, omvandlar h\u00f6gdimensionella spektraldata till f\u00e4rre komponenter som f\u00f6rklarar mest varians, vilket f\u00f6renklar analysen. Denna metod ignorerar dock rumsligt sammanhang, s\u00e5som arrangemanget av gr\u00f6dor p\u00e5 ett f\u00e4lt. Omv\u00e4nt lyfte rumsliga metoder, som matematiska morfologioperatorer, fram m\u00f6nster i gr\u00f6dornas fysiska utformning men f\u00f6rbisedde kritiska spektrala detaljer.<\/p>\n

Matematisk morfologi anv\u00e4nder operationer som dilatation och erosion f\u00f6r att extrahera former och strukturer fr\u00e5n bilder, s\u00e5som gr\u00e4nserna mellan f\u00e4lt. Med tiden f\u00f6rb\u00e4ttrade faltningsneurala n\u00e4tverk (CNN) klassificeringen genom att bearbeta b\u00e5da typerna av data.<\/p>\n

Deras fasta receptiva f\u00e4lt \u2013 den del av en bild som ett n\u00e4tverk kan "se" p\u00e5 en g\u00e5ng \u2013 begr\u00e4nsade dock deras f\u00f6rm\u00e5ga att f\u00e5nga l\u00e5ngsiktiga beroenden. Till exempel kan ett 3D-CNN ha sv\u00e5rt att skilja mellan tv\u00e5 sojab\u00f6nsorter med liknande spektralprofiler men olika tillv\u00e4xtm\u00f6nster \u00f6ver ett stort f\u00e4lt.<\/p>\n

Transformers, en typ av neuralt n\u00e4tverk som ursprungligen utformades f\u00f6r naturlig spr\u00e5kbehandling, erbj\u00f6d en l\u00f6sning p\u00e5 detta problem. Genom att anv\u00e4nda sj\u00e4lvuppm\u00e4rksamhetmekanismer utm\u00e4rker sig Transformers p\u00e5 att modellera globala relationer i data. Sj\u00e4lvuppm\u00e4rksamhet g\u00f6r det m\u00f6jligt f\u00f6r modellen att v\u00e4ga vikten av olika delar av en inmatningssekvens, vilket g\u00f6r att den kan fokusera p\u00e5 relevanta regioner (t.ex. ett kluster av sjuka v\u00e4xter) samtidigt som brus (t.ex. molnskuggor) ignoreras.<\/p>\n

\u00c4nd\u00e5 missar de ofta finkorniga lokala detaljer, s\u00e5som l\u00f6vkanter eller jordsprickor. Hybridmodeller som CTMixer f\u00f6rs\u00f6kte kombinera CNN och Transformers men gjorde det sekventiellt, och bearbetade lokala funktioner f\u00f6rst och globala funktioner senare. Denna metod ledde till ineffektiv sammanslagning av information och suboptimal prestanda i komplexa jordbruksmilj\u00f6er.<\/p>\n

Hur CMTNet fungerar: \u00d6verbrygga lokala och globala funktioner<\/h2>\n

CMTNet \u00f6vervinner dessa begr\u00e4nsningar genom en unik tredelad arkitektur utformad f\u00f6r att effektivt extrahera och sammanfoga spektral-rumsliga, lokala och globala funktioner.<\/p>\n

1.<\/strong> Den f\u00f6rsta komponenten, den modul f\u00f6r extraktion av spektral-rumsliga funktioner<\/strong>, bearbetar r\u00e5data i HSI med hj\u00e4lp av 3D- och 2D-faltningslager.<\/p>\n

3D-faltningsskikten analyserar b\u00e5de rumsliga (h\u00f6jd \u00d7 bredd) och spektrala (v\u00e5gl\u00e4ngds) dimensioner samtidigt och f\u00e5ngar m\u00f6nster som reflektansen av specifika v\u00e5gl\u00e4ngder \u00f6ver en gr\u00f6das tak. Till exempel kan en 3D-k\u00e4rna uppt\u00e4cka att frisk majs reflekterar mer n\u00e4ra-infrar\u00f6tt ljus i sina \u00f6vre blad j\u00e4mf\u00f6rt med de nedre.<\/p>\n

2D-lagren f\u00f6rfinar sedan dessa egenskaper och fokuserar p\u00e5 rumsliga detaljer som v\u00e4xternas arrangemang i ett f\u00e4lt. Denna tv\u00e5stegsprocess s\u00e4kerst\u00e4ller att b\u00e5de spektral m\u00e5ngfald (t.ex. klorofyllinneh\u00e5ll) och rumsligt sammanhang (t.ex. radavst\u00e5nd) bevaras.<\/p>\n

2.<\/strong> Den andra komponenten, den lokal-global funktionsutvinningsmodul<\/strong>, fungerar parallellt. En gren anv\u00e4nder CNN f\u00f6r att fokusera p\u00e5 lokala detaljer, s\u00e5som texturen p\u00e5 enskilda blad eller formen p\u00e5 jordfl\u00e4ckar. Dessa egenskaper \u00e4r avg\u00f6rande f\u00f6r att identifiera arter med liknande spektralprofiler, s\u00e5som olika sojab\u00f6nsorter.<\/p>\n

Den andra grenen anv\u00e4nder Transformers f\u00f6r att modellera globala samband, s\u00e5som hur gr\u00f6dor f\u00f6rdelas \u00f6ver stora omr\u00e5den eller hur skuggor fr\u00e5n n\u00e4rliggande tr\u00e4d p\u00e5verkar spektralv\u00e4rden. Genom att bearbeta dessa funktioner samtidigt snarare \u00e4n sekventiellt undviker CMTNet den informationsf\u00f6rlust som pl\u00e5gar tidigare hybridmodeller.<\/p>\n

Till exempel, medan CNN-grenen identifierar de taggiga kanterna p\u00e5 bomullsblad, inser Transformer-grenen att dessa blad \u00e4r en del av ett st\u00f6rre bomullsf\u00e4lt som gr\u00e4nsar till sesamplantor.<\/p>\n

3.<\/strong> Den tredje komponenten, den begr\u00e4nsningsmodul med flera utg\u00e5ngar<\/strong>, s\u00e4kerst\u00e4ller balanserad inl\u00e4rning \u00f6ver lokala, globala och sammanslagna funktioner. Under tr\u00e4ning till\u00e4mpas separata f\u00f6rlustfunktioner p\u00e5 varje typ av funktion, vilket tvingar n\u00e4tverket att f\u00f6rfina alla aspekter av sin f\u00f6rst\u00e5else.<\/p>\n

En f\u00f6rlustfunktion kvantifierar skillnaden mellan f\u00f6rutsp\u00e5dda och faktiska v\u00e4rden och v\u00e4gleder modellens justeringar. Till exempel kan f\u00f6rlusten f\u00f6r lokala egenskaper bestraffa modellen f\u00f6r felklassificering av bladkanter, medan den globala f\u00f6rlusten korrigerar fel i storskalig gr\u00f6df\u00f6rdelning.<\/p>\n

Dessa f\u00f6rluster kombineras med hj\u00e4lp av vikter som optimerats genom en slumpm\u00e4ssig s\u00f6kning \u2013 en teknik som testar olika viktkombinationer f\u00f6r att maximera noggrannheten. Denna process resulterar i en robust och anpassningsbar modell som kan hantera olika jordbruksscenarier.<\/p>\n

Utv\u00e4rdering av CMTNet-prestanda p\u00e5 hyperspektrala UAV-dataset<\/h2>\n

F\u00f6r att utv\u00e4rdera CMTNet testade forskare det p\u00e5 tre hyperspektrala datam\u00e4ngder fr\u00e5n Wuhan University, f\u00f6rv\u00e4rvade med dr\u00f6nare. Dessa datam\u00e4ngder anv\u00e4nds flitigt som riktm\u00e4rken inom fj\u00e4rranalys p\u00e5 grund av deras h\u00f6ga kvalitet och m\u00e5ngfald:<\/p>\n

    \n
  1. WHU-Hej-LongKou<\/strong>Denna dataupps\u00e4ttning t\u00e4cker 550 \u00d7 400 pixlar med 270 spektralband och en spatial uppl\u00f6sning p\u00e5 0,463 meter. En spatial uppl\u00f6sning p\u00e5 0,463 meter inneb\u00e4r att varje pixel representerar ett omr\u00e5de p\u00e5 0,463 m \u00d7 0,463 m p\u00e5 marken, vilket m\u00f6jligg\u00f6r identifiering av enskilda v\u00e4xter. Den inkluderar nio gr\u00f6dor, s\u00e5som majs, bomull och ris, med 1 019 tr\u00e4ningsprover och 203 523 testprover.<\/li>\n
  2. WHU-Hej-HanChuan<\/strong>Med en uppl\u00f6sning p\u00e5 1 217 \u00d7 303 pixlar och 0,109 meter visar denna dataupps\u00e4ttning 16 typer av markt\u00e4cke, inklusive jordgubbar, sojab\u00f6nor och plastfolie. Den h\u00f6gre uppl\u00f6sningen (0,109 m) m\u00f6jligg\u00f6r finare detaljer, s\u00e5som skillnaden mellan unga och mogna sojab\u00f6nsplantor. Tr\u00e4nings- och testproverna uppgick till totalt 1 289 respektive 256 241.<\/li>\n
  3. WHU-Hej-HongHu<\/strong>Med 940 \u00d7 475 pixlar och 270 band inneh\u00e5ller denna h\u00f6guppl\u00f6sta (0,043 meter) dataupps\u00e4ttning 22 klasser, s\u00e5som bomulls-, raps- och vitl\u00f6ksskott. Vid 0,043 m uppl\u00f6sning \u00e4r enskilda blad och jordsprickor synliga, vilket g\u00f6r den idealisk f\u00f6r finkornig klassificering. Den inneh\u00e5ller 1 925 tr\u00e4ningsprover och 384 678 testprover.<\/li>\n<\/ol>\n

    \"J\u00e4mf\u00f6relse<\/p>\n

    Modellen tr\u00e4nades p\u00e5 NVIDIA TITAN Xp GPU:er med PyTorch, med en inl\u00e4rningshastighet p\u00e5 0,001 och en batchstorlek p\u00e5 100. En inl\u00e4rningshastighet avg\u00f6r hur mycket modellen justerar sina parametrar under tr\u00e4ningen \u2013 f\u00f6r h\u00f6g kan den \u00f6verskrida optimala v\u00e4rden; f\u00f6r l\u00e5g blir tr\u00e4ningen tr\u00f6g.<\/p>\n

    Varje experiment upprepades tio g\u00e5nger f\u00f6r att s\u00e4kerst\u00e4lla tillf\u00f6rlitlighet, och inmatningsfl\u00e4ckar \u2013 sm\u00e5 segment av hela bilden \u2013 optimerades till 13 \u00d7 13 pixlar genom rutn\u00e4tss\u00f6kning, en metod som testar olika patchstorlekar f\u00f6r att hitta den mest effektiva.<\/p>\n

    CMTNet uppn\u00e5r toppmodern noggrannhet i klassificering av gr\u00f6dor<\/h2>\n

    CMTNet uppn\u00e5dde anm\u00e4rkningsv\u00e4rda resultat \u00f6ver alla datam\u00e4ngder och \u00f6vertr\u00e4ffade befintliga metoder b\u00e5de vad g\u00e4ller total noggrannhet (OA) och klassspecifik prestanda. OA m\u00e4ter andelen korrekt klassificerade pixlar \u00f6ver alla klasser, medan genomsnittlig noggrannhet (AA) ber\u00e4knar den genomsnittliga noggrannheten per klass och \u00e5tg\u00e4rdar obalanser.<\/p>\n

    P\u00e5 WHU-Hi-LongKou-datasetet uppn\u00e5dde CMTNet en OA p\u00e5 99,58%, vilket \u00f6vertr\u00e4ffade CTMixer med 0,19%. F\u00f6r utmanande klasser med begr\u00e4nsad tr\u00e4ningsdata, s\u00e5som bomull (41 prover), n\u00e5dde CMTNet fortfarande en noggrannhet p\u00e5 99,53%. P\u00e5 liknande s\u00e4tt f\u00f6rb\u00e4ttrade WHU-Hi-HanChuan-datasetet noggrannheten f\u00f6r vattenmelon (22 prover) fr\u00e5n 82,42% till 96,11%, vilket demonstrerar dess f\u00f6rm\u00e5ga att hantera obalanserade data genom effektiv funktionsfusion.<\/p>\n

    Visuella j\u00e4mf\u00f6relser av klassificeringskartor avsl\u00f6jade f\u00e4rre fragmenterade fl\u00e4ckar och j\u00e4mnare gr\u00e4nser mellan f\u00e4lt j\u00e4mf\u00f6rt med modeller som 3D-CNN och Vision Transformer (ViT). Till exempel, i den skuggben\u00e4gna WHU-Hi-HanChuan-dataupps\u00e4ttningen minimerade CMTNet fel orsakade av l\u00e5ga solvinklar, medan ResNet felklassificerade sojab\u00f6nor som gr\u00e5a tak.<\/p>\n

    \"CMTNets<\/p>\n

    Skuggor utg\u00f6r en unik utmaning eftersom de f\u00f6r\u00e4ndrar spektrala signaturer \u2013 en sojab\u00f6nsplanta i skugga kan reflektera mindre n\u00e4ra-infrar\u00f6tt ljus, vilket liknar icke-vegetation. Genom att utnyttja det globala sammanhanget ins\u00e5g CMTNet att dessa skuggade v\u00e4xter var en del av ett st\u00f6rre sojab\u00f6nsf\u00e4lt, vilket minskade fel.<\/p>\n

    P\u00e5 WHU-Hi-HongHu-datasetet utm\u00e4rkte sig modellen i att s\u00e4rskilja spektralt liknande gr\u00f6dor, s\u00e5som olika k\u00e5lsorter, och uppn\u00e5dde en noggrannhet p\u00e5 96,54% f\u00f6r\u00a0Brassica parachinensis<\/em>.<\/p>\n

    Ablationsstudier \u2013 experiment som tar bort komponenter f\u00f6r att bed\u00f6ma deras inverkan \u2013 bekr\u00e4ftade vikten av varje modul. Genom att enbart l\u00e4gga till multi-output constraint-modulen \u00f6kade OA med 1,52% p\u00e5 WHU-Hi-HongHu, vilket belyser dess roll i att f\u00f6rfina funktionsfusionen. Utan denna modul kombinerades lokala och globala funktioner slumpm\u00e4ssigt, vilket ledde till inkonsekventa klassificeringar.<\/p>\n

    Ber\u00e4kningsm\u00e4ssiga avv\u00e4gningar och praktiska \u00f6verv\u00e4ganden<\/h2>\n

    Medan CMTNets noggrannhet \u00e4r o\u00f6vertr\u00e4ffad, \u00e4r dess ber\u00e4kningskostnad h\u00f6gre \u00e4n traditionella metoders. Tr\u00e4ning p\u00e5 WHU-Hi-HongHu-datasetet tog 1 885 sekunder, j\u00e4mf\u00f6rt med 74 sekunder f\u00f6r Random Forest (RF), en maskininl\u00e4rningsalgoritm som bygger beslutstr\u00e4d under tr\u00e4ning.<\/p>\n

    Denna avv\u00e4gning \u00e4r dock ber\u00e4ttigad inom precisionsjordbruk, d\u00e4r noggrannhet direkt p\u00e5verkar avkastningsprognoser och resursallokering. Till exempel kan felklassificering av en sjuk gr\u00f6da som frisk leda till okontrollerade skadedjursutbrott, vilket \u00f6del\u00e4gger hela f\u00e4lt.<\/p>\n

    F\u00f6r realtidsapplikationer skulle framtida arbete kunna utforska modellkomprimeringstekniker, s\u00e5som att besk\u00e4ra redundanta neuroner eller kvantisera vikter (vilket minskar numerisk precision), f\u00f6r att minska k\u00f6rtiden utan att offra prestanda. Besk\u00e4rning tar bort mindre viktiga kopplingar fr\u00e5n det neurala n\u00e4tverket, ungef\u00e4r som att trimma grenar fr\u00e5n ett tr\u00e4d f\u00f6r att f\u00f6rb\u00e4ttra dess form, medan kvantisering f\u00f6renklar numeriska ber\u00e4kningar och p\u00e5skyndar bearbetningen.<\/p>\n

    Framtiden f\u00f6r hyperspektral gr\u00f6dklassificering med CMTNet<\/h2>\n

    Trots sin framg\u00e5ng har CMTNet begr\u00e4nsningar. Prestandan sjunker n\u00e5got i starkt skuggade omr\u00e5den, vilket framg\u00e5r av WHU-Hi-HanChuan-datasetet (97.29% OA vs. 99.58% i v\u00e4l upplysta LongKou). Skuggor komplicerar klassificeringen eftersom de minskar intensiteten hos det reflekterade ljuset och f\u00f6r\u00e4ndrar spektralprofiler.<\/p>\n

    Dessutom halkar klasser med extremt sm\u00e5 tr\u00e4ningsurval, som smalbladiga sojab\u00f6nor (20 prover), efter de med rikligt med data. Sm\u00e5 urvalsstorlekar begr\u00e4nsar modellens f\u00f6rm\u00e5ga att l\u00e4ra sig olika variationer, s\u00e5som skillnader i bladform p\u00e5 grund av jordkvalitet.<\/p>\n

    Framtida forskning skulle kunna integrera multimodala data, s\u00e5som LiDAR-h\u00f6jdkartor eller v\u00e4rmebilder, f\u00f6r att f\u00f6rb\u00e4ttra motst\u00e5ndskraften mot skuggor och ocklusioner. LiDAR (Light Detection and Ranging) anv\u00e4nder laserpulser f\u00f6r att skapa 3D-terr\u00e4ngmodeller, vilket kan hj\u00e4lpa till att skilja gr\u00f6dor fr\u00e5n skuggor genom att analysera h\u00f6jdskillnader.<\/p>\n

    Dessutom f\u00e5ngar v\u00e4rmebilder v\u00e4rmesignaturer, vilket ger ytterligare ledtr\u00e5dar om v\u00e4xternas h\u00e4lsa \u2013 stressade gr\u00f6dor har ofta h\u00f6gre temperaturer i tr\u00e4dkronorna p\u00e5 grund av minskad transpiration. Semi\u00f6vervakade inl\u00e4rningstekniker, som utnyttjar om\u00e4rkta data (t.ex. UAV-bilder utan manuella annoteringar), kan ocks\u00e5 f\u00f6rb\u00e4ttra prestandan f\u00f6r s\u00e4llsynta gr\u00f6dor.<\/p>\n

    Genom att anv\u00e4nda konsistensregularisering \u2013 att tr\u00e4na modellen att producera stabila f\u00f6ruts\u00e4gelser \u00f6ver n\u00e5got f\u00f6r\u00e4ndrade versioner av samma bild \u2013 kan forskare utnyttja om\u00e4rkta data f\u00f6r att f\u00f6rb\u00e4ttra generaliseringen.<\/p>\n

    Slutligen skulle implementering av CMTNet p\u00e5 edge-enheter, som dr\u00f6nare utrustade med inbyggda GPU:er, kunna m\u00f6jligg\u00f6ra realtids\u00f6vervakning i avl\u00e4gsna f\u00e4lt. Edge-implementering minskar beroendet av molntj\u00e4nster, vilket minimerar latens och data\u00f6verf\u00f6ringskostnader. Detta kr\u00e4ver dock att modellen optimeras f\u00f6r begr\u00e4nsat minne och processorkraft, potentiellt genom l\u00e4ttviktsarkitekturer som MobileNet eller kunskapsdestillation, d\u00e4r en mindre "student"-modell efterliknar en st\u00f6rre "l\u00e4rar"-modell.<\/p>\n

    Slutsats<\/h2>\n

    CMTNet representerar ett betydande steg fram\u00e5t inom hyperspektral gr\u00f6dklassificering. Genom att harmonisera CNN och transformatorer tar det itu med l\u00e5ngvariga utmaningar inom extraktion och fusion av egenskaper, vilket ger jordbrukare och agronomer ett kraftfullt verktyg f\u00f6r precisionsjordbruk.<\/p>\n

    Till\u00e4mpningar str\u00e4cker sig fr\u00e5n sjukdomsdetektering i realtid till optimering av bevattningsscheman, vilka alla \u00e4r avg\u00f6rande f\u00f6r h\u00e5llbart jordbruk mitt i klimatf\u00f6r\u00e4ndringar och befolkningstillv\u00e4xt. I takt med att UAV-teknik blir mer tillg\u00e4nglig kommer modeller som CMTNet att spela en avg\u00f6rande roll i den globala livsmedelss\u00e4kerheten.<\/p>\n

    Framtida framsteg, s\u00e5som l\u00e4ttare arkitekturer och multimodal datafusion, skulle kunna f\u00f6rb\u00e4ttra deras praktiska anv\u00e4ndbarhet ytterligare. Med fortsatt innovation skulle CMTNet kunna bli en h\u00f6rnsten i smarta jordbrukssystem v\u00e4rlden \u00f6ver, vilket s\u00e4kerst\u00e4ller effektiv markanv\u00e4ndning och motst\u00e5ndskraftig livsmedelsproduktion f\u00f6r kommande generationer.<\/p>\n

    H\u00e4nvisning: <\/strong>Guo, X., Feng, Q. & Guo, F. CMTNet: ett hybrid CNN-transformatorn\u00e4tverk f\u00f6r UAV-baserad hyperspektral gr\u00f6dklassificering inom precisionsjordbruk. Sci Rep 15, 12383 (2025). https:\/\/doi.org\/10.1038\/s41598-025-97052-w<\/a><\/p>","protected":false},"excerpt":{"rendered":"

    Noggrann klassificering av gr\u00f6dor \u00e4r avg\u00f6rande f\u00f6r modernt precisionsjordbruk, vilket g\u00f6r det m\u00f6jligt f\u00f6r jordbrukare att \u00f6vervaka gr\u00f6dornas h\u00e4lsa, f\u00f6ruts\u00e4ga avkastning och f\u00f6rdela resurser effektivt. Traditionella metoder \u00e4r dock ofta\u2026<\/p>","protected":false},"author":210157960,"featured_media":11556,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_coblocks_attr":"","_coblocks_dimensions":"","_coblocks_responsive_height":"","_coblocks_accordion_ie_support":"","_eb_attr":"","content-type":"","_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_wpcom_ai_launchpad_first_post":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"{title}\n\n{excerpt}\n\n{url}","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"_wpas_customize_per_network":false,"jetpack_post_was_ever_published":false},"categories":[1657,1377],"tags":[],"class_list":["post-11549","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-precision-farming","category-crop-monitoring"],"acf":[],"yoast_head":"\nCMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification - GeoPard Agriculture<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/geopard.tech\/swe\/blogg\/cmtnet-omdefinierar-precisionsjordbruk-genom-att-overtraffa-traditionell-grodklassificering\/\" \/>\n<meta property=\"og:locale\" content=\"sv_SE\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification - GeoPard Agriculture\" \/>\n<meta property=\"og:description\" content=\"Accurate crop classification is essential for modern precision agriculture, enabling farmers to monitor crop health, predict yields, and allocate resources efficiently. Traditional methods, however, often...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/geopard.tech\/swe\/blogg\/cmtnet-omdefinierar-precisionsjordbruk-genom-att-overtraffa-traditionell-grodklassificering\/\" \/>\n<meta property=\"og:site_name\" content=\"GeoPard - Precision agriculture Mapping software\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/geopardAgriculture\/\" \/>\n<meta property=\"article:published_time\" content=\"2025-05-04T19:00:25+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification-1024x576.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1024\" \/>\n\t<meta property=\"og:image:height\" content=\"576\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"dementievgeopard\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@geopardagri\" \/>\n<meta name=\"twitter:site\" content=\"@geopardagri\" \/>\n<meta name=\"twitter:label1\" content=\"Skriven av\" \/>\n\t<meta name=\"twitter:data1\" content=\"dementievgeopard\" \/>\n\t<meta name=\"twitter:label2\" content=\"Ber\u00e4knad l\u00e4stid\" \/>\n\t<meta name=\"twitter:data2\" content=\"10 minuter\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/\"},\"author\":{\"name\":\"dementievgeopard\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/person\\\/dd217733c742620adc57befbbcd84a8a\"},\"headline\":\"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification\",\"datePublished\":\"2025-05-04T19:00:25+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/\"},\"wordCount\":2064,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1\",\"articleSection\":[\"Precision Farming\",\"Crop monitoring\"],\"inLanguage\":\"sv-SE\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/\",\"url\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/\",\"name\":\"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification - GeoPard Agriculture\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1\",\"datePublished\":\"2025-05-04T19:00:25+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#breadcrumb\"},\"inLanguage\":\"sv-SE\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"sv-SE\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#primaryimage\",\"url\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1\",\"contentUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1\",\"width\":3600,\"height\":2025},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/geopard.tech\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#website\",\"url\":\"https:\\\/\\\/geopard.tech\\\/\",\"name\":\"GeoPard - Precision agriculture software\",\"description\":\"Precision agriculture Mapping software\",\"publisher\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/#organization\"},\"alternateName\":\"GeoPard\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/geopard.tech\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"sv-SE\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#organization\",\"name\":\"GeoPard Agriculture\",\"alternateName\":\"GeoPard\",\"url\":\"https:\\\/\\\/geopard.tech\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"sv-SE\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2026\\\/07\\\/geopard_logo_transparentbackground.png?fit=512%2C68&ssl=1\",\"contentUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2026\\\/07\\\/geopard_logo_transparentbackground.png?fit=512%2C68&ssl=1\",\"width\":512,\"height\":68,\"caption\":\"GeoPard Agriculture\"},\"image\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/geopardAgriculture\\\/\",\"https:\\\/\\\/x.com\\\/geopardagri\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/geopard-agriculture\\\/\",\"https:\\\/\\\/www.instagram.com\\\/geopardagriculture\\\/\",\"https:\\\/\\\/www.youtube.com\\\/channel\\\/UCiaPGLAhRPNh-s85dXdC-Sw\",\"https:\\\/\\\/www.g2.com\\\/products\\\/geopard-agriculture-precision-farming-software\\\/reviews\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/person\\\/dd217733c742620adc57befbbcd84a8a\",\"name\":\"dementievgeopard\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"sv-SE\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/7c09adac9c20b00454199df5a41ea67c812a6c97e948bf109899836ddde31354?s=96&d=identicon&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/7c09adac9c20b00454199df5a41ea67c812a6c97e948bf109899836ddde31354?s=96&d=identicon&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/7c09adac9c20b00454199df5a41ea67c812a6c97e948bf109899836ddde31354?s=96&d=identicon&r=g\",\"caption\":\"dementievgeopard\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"CMTNet omdefinierar precisionsjordbruk genom att \u00f6vertr\u00e4ffa traditionell gr\u00f6dklassificering - GeoPard Agriculture","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/geopard.tech\/swe\/blogg\/cmtnet-omdefinierar-precisionsjordbruk-genom-att-overtraffa-traditionell-grodklassificering\/","og_locale":"sv_SE","og_type":"article","og_title":"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification - GeoPard Agriculture","og_description":"Accurate crop classification is essential for modern precision agriculture, enabling farmers to monitor crop health, predict yields, and allocate resources efficiently. Traditional methods, however, often...","og_url":"https:\/\/geopard.tech\/swe\/blogg\/cmtnet-omdefinierar-precisionsjordbruk-genom-att-overtraffa-traditionell-grodklassificering\/","og_site_name":"GeoPard - Precision agriculture Mapping software","article_publisher":"https:\/\/www.facebook.com\/geopardAgriculture\/","article_published_time":"2025-05-04T19:00:25+00:00","og_image":[{"width":1024,"height":576,"url":"https:\/\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification-1024x576.png","type":"image\/png"}],"author":"dementievgeopard","twitter_card":"summary_large_image","twitter_creator":"@geopardagri","twitter_site":"@geopardagri","twitter_misc":{"Skriven av":"dementievgeopard","Ber\u00e4knad l\u00e4stid":"10 minuter"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#article","isPartOf":{"@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/"},"author":{"name":"dementievgeopard","@id":"https:\/\/geopard.tech\/#\/schema\/person\/dd217733c742620adc57befbbcd84a8a"},"headline":"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification","datePublished":"2025-05-04T19:00:25+00:00","mainEntityOfPage":{"@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/"},"wordCount":2064,"commentCount":0,"publisher":{"@id":"https:\/\/geopard.tech\/#organization"},"image":{"@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#primaryimage"},"thumbnailUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1","articleSection":["Precision Farming","Crop monitoring"],"inLanguage":"sv-SE","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/","url":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/","name":"CMTNet omdefinierar precisionsjordbruk genom att \u00f6vertr\u00e4ffa traditionell gr\u00f6dklassificering - GeoPard Agriculture","isPartOf":{"@id":"https:\/\/geopard.tech\/#website"},"primaryImageOfPage":{"@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#primaryimage"},"image":{"@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#primaryimage"},"thumbnailUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1","datePublished":"2025-05-04T19:00:25+00:00","breadcrumb":{"@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#breadcrumb"},"inLanguage":"sv-SE","potentialAction":[{"@type":"ReadAction","target":["https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/"]}]},{"@type":"ImageObject","inLanguage":"sv-SE","@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#primaryimage","url":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1","contentUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1","width":3600,"height":2025},{"@type":"BreadcrumbList","@id":"https:\/\/geopard.tech\/blog\/cmtnet-redefines-precision-agriculture-by-outperforming-traditional-crop-classification\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/geopard.tech\/"},{"@type":"ListItem","position":2,"name":"CMTNet Redefines Precision Agriculture By Outperforming Traditional Crop Classification"}]},{"@type":"WebSite","@id":"https:\/\/geopard.tech\/#website","url":"https:\/\/geopard.tech\/","name":"GeoPard - Precisionsjordbruksmjukvara","description":"Precisionsjordbruk kartprogramvara","publisher":{"@id":"https:\/\/geopard.tech\/#organization"},"alternateName":"GeoPard","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/geopard.tech\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"sv-SE"},{"@type":"Organization","@id":"https:\/\/geopard.tech\/#organization","name":"GeoPard Jordbruk","alternateName":"GeoPard","url":"https:\/\/geopard.tech\/","logo":{"@type":"ImageObject","inLanguage":"sv-SE","@id":"https:\/\/geopard.tech\/#\/schema\/logo\/image\/","url":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2026\/07\/geopard_logo_transparentbackground.png?fit=512%2C68&ssl=1","contentUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2026\/07\/geopard_logo_transparentbackground.png?fit=512%2C68&ssl=1","width":512,"height":68,"caption":"GeoPard Agriculture"},"image":{"@id":"https:\/\/geopard.tech\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/geopardAgriculture\/","https:\/\/x.com\/geopardagri","https:\/\/www.linkedin.com\/company\/geopard-agriculture\/","https:\/\/www.instagram.com\/geopardagriculture\/","https:\/\/www.youtube.com\/channel\/UCiaPGLAhRPNh-s85dXdC-Sw","https:\/\/www.g2.com\/products\/geopard-agriculture-precision-farming-software\/reviews"]},{"@type":"Person","@id":"https:\/\/geopard.tech\/#\/schema\/person\/dd217733c742620adc57befbbcd84a8a","name":"Demens","image":{"@type":"ImageObject","inLanguage":"sv-SE","@id":"https:\/\/secure.gravatar.com\/avatar\/7c09adac9c20b00454199df5a41ea67c812a6c97e948bf109899836ddde31354?s=96&d=identicon&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/7c09adac9c20b00454199df5a41ea67c812a6c97e948bf109899836ddde31354?s=96&d=identicon&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/7c09adac9c20b00454199df5a41ea67c812a6c97e948bf109899836ddde31354?s=96&d=identicon&r=g","caption":"dementievgeopard"}}]}},"jetpack_publicize_connections":[],"jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/pdiCPa-30h","jetpack_featured_media_url":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/CMTNet-Redefines-Precision-Agriculture-By-Outperforming-Traditional-Crop-Classification.png?fit=3600%2C2025&ssl=1","_links":{"self":[{"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/posts\/11549","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/users\/210157960"}],"replies":[{"embeddable":true,"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/comments?post=11549"}],"version-history":[{"count":0,"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/posts\/11549\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/media\/11556"}],"wp:attachment":[{"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/media?parent=11549"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/categories?post=11549"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/geopard.tech\/swe\/wp-json\/wp\/v2\/tags?post=11549"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}