{"id":11559,"date":"2025-05-04T22:58:53","date_gmt":"2025-05-04T20:58:53","guid":{"rendered":"https:\/\/geopard.tech\/?p=11559"},"modified":"2025-05-04T23:00:29","modified_gmt":"2025-05-04T21:00:29","slug":"uzgoj-jecma-dobiva-poticaj-laganom-detekcijom-yolov5","status":"publish","type":"post","link":"https:\/\/geopard.tech\/hr\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/","title":{"rendered":"Uzgoj je\u010dma potpomognut laganom YOLOv5 detekcijom"},"content":{"rendered":"

Je\u010dam iz visokogorskog podru\u010dja, otporna \u017eitarica koja se uzgaja u visokogorskim regijama kineske visoravni Qinghai-Tibet, igra klju\u010dnu ulogu u lokalnoj sigurnosti hrane i gospodarskoj stabilnosti. Znanstveno poznat kao\u00a0Hordeum vulgare<\/em>\u00a0L., ova kultura uspijeva u ekstremnim uvjetima - rijetkom zraku, niskoj razini kisika i prosje\u010dnoj godi\u0161njoj temperaturi od 6,3 \u00b0C - \u0161to je \u010dini neophodnom za zajednice u te\u0161kim okru\u017eenjima.<\/p>\n

S vi\u0161e od 270.000 hektara posve\u0107enih uzgoju u Kini, prvenstveno u autonomnoj regiji Xizang, visokogorski je\u010dam \u010dini vi\u0161e od polovice zasa\u0111ene povr\u0161ine regije i preko 70% ukupne proizvodnje \u017eitarica. To\u010dno pra\u0107enje gusto\u0107e je\u010dma - broja biljaka ili klasova po jedinici povr\u0161ine - klju\u010dno je za optimizaciju poljoprivrednih praksi, poput navodnjavanja i gnojidbe, te predvi\u0111anje prinosa.<\/p>\n

Me\u0111utim, tradicionalne metode poput ru\u010dnog uzorkovanja ili satelitskog snimanja pokazale su se neu\u010dinkovitima, radno intenzivnima ili nedovoljno detaljnima. Kako bi se suo\u010dili s tim izazovima, istra\u017eiva\u010di sa Sveu\u010dili\u0161ta za poljoprivredu i \u0161umarstvo Fujian i Tehnolo\u0161kog sveu\u010dili\u0161ta Chengdu razvili su inovativni model umjetne inteligencije temeljen na YOLOv5, vrhunskom algoritmu za detekciju objekata.<\/p>\n

Njihov rad, objavljen u\u00a0Metode sadnje<\/em>\u00a0(2025.) postigli su izvanredne rezultate, uklju\u010duju\u0107i srednju prosje\u010dnu preciznost (mAP) od 93,1% - metriku koja mjeri ukupnu to\u010dnost detekcije - i smanjenje ra\u010dunalnih tro\u0161kova od 75,6%, \u0161to ga \u010dini prikladnim za primjenu dronova u stvarnom vremenu.<\/p>\n

Izazovi i inovacije u pra\u0107enju usjeva<\/h2>\n

Va\u017enost visokogorskog je\u010dma nadilazi njegovu ulogu kao izvora hrane. Samo u 2022. godini, grad Rikaze, glavna regija za proizvodnju je\u010dma, po\u017enjeo je 408.900 tona je\u010dma na 60.000 hektara, \u0161to \u010dini gotovo polovicu ukupne proizvodnje \u017eitarica u Tibetu.<\/p>\n

Unato\u010d kulturnom i ekonomskom zna\u010daju, procjena prinosa je\u010dma dugo je bila izazovna. Tradicionalne metode, poput ru\u010dnog brojanja ili satelitskih snimaka, ili su previ\u0161e radno intenzivne ili im nedostaje rezolucija potrebna za otkrivanje pojedina\u010dnih klasova je\u010dma - dijela biljke koji nosi zrno, a koji su \u010desto \u0161iroki samo 2-3 centimetra.<\/p>\n

Ru\u010dno uzorkovanje zahtijeva od poljoprivrednika da fizi\u010dki pregledaju dijelove polja - proces koji je spor, subjektivan i neprakti\u010dan za velike farme. Satelitske snimke, iako korisne za \u0161iroka promatranja, imaju problema s niskom rezolucijom (\u010desto 10-30 metara po pikselu) i \u010destim vremenskim poreme\u0107ajima, poput naoblake u planinskim regijama poput Tibeta.<\/p>\n

Kako bi prevladali ta ograni\u010denja, istra\u017eiva\u010di su se okrenuli bespilotnim letjelicama (UAV) ili dronovima, opremljenim kamerama od 20 megapiksela. Ove su dronove snimile 501 sliku visoke rezolucije polja je\u010dma u gradu Rikaze tijekom dvije kriti\u010dne faze rasta: faze rasta u kolovozu 2022., koju karakteriziraju zeleni, razvijaju\u0107i se klasovi, i faze sazrijevanja u kolovozu 2023., koju obilje\u017eavaju zlatno-\u017euti, klasovi spremni za \u017eetvu.<\/p>\n

\"Pra\u0107enje<\/p>\n

Me\u0111utim, analiza ovih slika predstavljala je izazove, uklju\u010duju\u0107i zamu\u0107ene rubove uzrokovane kretanjem drona, malu veli\u010dinu klasova je\u010dma na zra\u010dnim snimkama i preklapaju\u0107e klasove u gusto zasa\u0111enim poljima.<\/p>\n

Kako bi rije\u0161ili ove probleme, istra\u017eiva\u010di su prethodno obradili slike dijeljenjem svake slike visoke rezolucije na 35 manjih podslika i filtriranjem mutnih rubova, \u0161to je rezultiralo s 2970 visokokvalitetnih podslika za obuku. Ovaj korak prethodne obrade osigurao je da se model usredoto\u010di na jasne, uporabne podatke, izbjegavaju\u0107i ometanja od podru\u010dja niske kvalitete.<\/p>\n

Tehni\u010dki napredak u detekciji objekata<\/h2>\n

Sredi\u0161nji dio ovog istra\u017eivanja je YOLOv5 algoritam (You Only Look Once verzija 5), jednostupanjski model detekcije objekata poznat po svojoj brzini i modularnom dizajnu. Za razliku od starijih dvostupanjskih modela poput Faster R-CNN-a, koji prvo identificiraju podru\u010dja interesa, a zatim klasificiraju objekte, YOLOv5 detekciju izvodi u jednom prolazu, \u0161to je \u010dini znatno br\u017eom.<\/p>\n

Osnovni YOLOv5n model, s 1,76 milijuna parametara (konfigurabilne komponente AI modela) i 4,1 milijardom FLOP-ova (operacije s pomi\u010dnim zarezom, mjera ra\u010dunalne slo\u017eenosti), ve\u0107 je bio u\u010dinkovit. Me\u0111utim, otkrivanje sitnih, preklapaju\u0107ih skokova je\u010dma zahtijevalo je daljnju optimizaciju.<\/p>\n

Istra\u017eiva\u010dki tim uveo je tri klju\u010dna pobolj\u0161anja modela: dubinski odvojivu konvoluciju (DSConv), duhovnu konvoluciju (GhostConv) i konvolucijski blokovni modul pa\u017enje (CBAM).<\/p>\n

Dubinski odvojiva konvolucija (DSConv) smanjuje ra\u010dunalne tro\u0161kove dijeljenjem standardnog procesa konvolucije - matemati\u010dke operacije koja izdvaja zna\u010dajke iz slika - u dva koraka. Prvo, dubinski konvolucija primjenjuje filtere na pojedina\u010dne kanale boja (npr. crvenu, zelenu, plavu), analiziraju\u0107i svaki kanal zasebno.<\/p>\n

Nakon toga slijedi to\u010dkasta konvolucija, koja kombinira rezultate preko kanala koriste\u0107i 1\u00d71 kernele. Ovaj pristup smanjuje broj parametara do 75%.<\/p>\n

\"Redukcija<\/p>\n

Na primjer, tradicionalna 3\u00d73 konvolucija sa 64 ulazna i 128 izlaznih kanala zahtijeva 73 728 parametara, dok DSConv to smanjuje na samo 8 768 - smanjenje od 88%. Ova u\u010dinkovitost je klju\u010dna za implementaciju modela na dronovima ili mobilnim ure\u0111ajima s ograni\u010denom procesorskom snagom.<\/p>\n

Ghost konvolucija (GhostConv) dodatno olak\u0161ava model generiranjem dodatnih mapa zna\u010dajki - pojednostavljenih prikaza uzoraka slike - kroz jednostavne linearne operacije, poput rotacije ili skaliranja, umjesto konvolucija koje zahtijevaju puno resursa.<\/p>\n

Tradicionalni konvolucijski slojevi stvaraju redundantne zna\u010dajke, tro\u0161e\u0107i ra\u010dunalne resurse. GhostConv to rje\u0161ava stvaranjem "fantomskih" zna\u010dajki iz postoje\u0107ih, u\u010dinkovito prepolovljuju\u0107i parametre u odre\u0111enim slojevima.<\/p>\n

Na primjer, sloj sa 64 ulazna i 128 izlaznih kanala tradicionalno bi zahtijevao\u00a073.728 parametara<\/strong>, ali GhostConv to svodi na\u00a036,864<\/strong>\u00a0uz odr\u017eavanje to\u010dnosti. Ova tehnika je posebno korisna za otkrivanje malih objekata poput klasova je\u010dma, gdje je ra\u010dunalna u\u010dinkovitost od najve\u0107e va\u017enosti.<\/p>\n

Modul konvolucijske blokovne pa\u017enje (CBAM) integriran je kako bi se modelu pomoglo da se usredoto\u010di na kriti\u010dne zna\u010dajke, \u010dak i u pretrpanim okru\u017eenjima. Mehanizmi pa\u017enje, inspirirani ljudskim vizualnim sustavima, omogu\u0107uju modelima umjetne inteligencije da daju prioritet va\u017enim dijelovima slike.<\/p>\n

CBAM koristi dvije vrste pa\u017enje: pa\u017enju na kanal, koja identificira va\u017ene kanale boja (npr. zelenu za rastu\u0107e klasove), i prostornu pa\u017enju, koja isti\u010de klju\u010dna podru\u010dja unutar slike (npr. nakupine klasova). Zamjenom standardnih modula s DSConv i GhostConv te uklju\u010divanjem CBAM-a, istra\u017eiva\u010di su stvorili precizniji model prilago\u0111en detekciji je\u010dma.<\/p>\n

Provedba i rezultati<\/h2>\n

Za treniranje modela, istra\u017eiva\u010di su ru\u010dno ozna\u010dili 135 originalnih slika pomo\u0107u grani\u010dnih okvira - pravokutnih okvira koji ozna\u010davaju polo\u017eaj klasova je\u010dma - kategoriziraju\u0107i klasove u faze rasta i sazrijevanja. Tehnike pro\u0161irenja podataka - uklju\u010duju\u0107i rotaciju, ubrizgavanje \u0161uma, okluziju i izo\u0161travanje - pro\u0161irile su skup podataka na 2970 slika, pobolj\u0161avaju\u0107i sposobnost modela da generalizira u razli\u010ditim terenskim uvjetima.<\/p>\n

Na primjer, rotiranje slika za 90\u00b0, 180\u00b0 ili 270\u00b0 pomoglo je modelu da prepozna \u0161iljke iz razli\u010ditih kutova, a istovremeno je dodao \u0161um simuliran nesavr\u0161enostima iz stvarnog svijeta poput pra\u0161ine ili sjena. Skup podataka podijeljen je u skup za u\u010denje (80%) i skup za validaciju (20%), osiguravaju\u0107i robusnu evaluaciju.<\/p>\n

Trening se odvijao na visokou\u010dinkovitom sustavu s AMD Ryzen 7 CPU-om, NVIDIA RTX 4060 GPU-om i 64 GB RAM-a, koriste\u0107i PyTorch framework - popularan alat za duboko u\u010denje. Pomno su pra\u0107eni preko 300 epoha treninga (potpunih prolaza kroz skup podataka), preciznost modela (to\u010dnost ispravnih detekcija), prisje\u0107anje (sposobnost pronala\u017eenja svih relevantnih skokova) i gubitak (stopa pogre\u0161aka).<\/p>\n

Rezultati su bili zapanjuju\u0107i. Pobolj\u0161ani YOLOv5 model postigao je preciznost od 92,2% (u odnosu na 89,1% u po\u010detnom modelu) i pouzdanost od 86,2% (u odnosu na 83,1%), nadma\u0161iv\u0161i po\u010detni YOLOv5n za 3,1% u obje metrike. Njegova srednja prosje\u010dna preciznost (mAP) - sveobuhvatna metrika koja usrednjava to\u010dnost detekcije u svim kategorijama - dosegla je 93,1%, s pojedina\u010dnim rezultatima od 92,7% za skokove u fazi rasta i 93,5% za skokove u fazi sazrijevanja.<\/p>\n

\"Rezultati<\/p>\n

Jednako impresivna bila je i njegova ra\u010dunalna u\u010dinkovitost: parametri modela pali su za 70,6% na 1,2 milijuna, a FLOP-ovi su se smanjili za 75,6% na 3,1 milijardu. Usporedne analize s vode\u0107im modelima poput Faster R-CNN i YOLOv8n istaknule su njegovu superiornost.<\/p>\n

Iako je YOLOv8n postigao ne\u0161to ve\u0107i mAP (93,8%), njegovi parametri (3,0 milijuna) i FLOP-ovi (8,1 milijarda) bili su 2,5x odnosno 2,6x ve\u0107i, \u0161to predlo\u017eeni model \u010dini daleko u\u010dinkovitijim za primjene u stvarnom vremenu.<\/p>\n

Vizualne usporedbe naglasile su ovaj napredak. Na slikama u fazi rasta, pobolj\u0161ani model otkrio je 41 \u0161iljak u usporedbi s 28 na po\u010detnoj liniji. Tijekom sazrijevanja identificirao je 3 \u0161iljka u usporedbi s 2 na po\u010detnoj liniji, s manje propu\u0161tenih detekcija (ozna\u010denih naran\u010dastim strelicama) i la\u017eno pozitivnih rezultata (ozna\u010denih ljubi\u010dastim strelicama).<\/p>\n

Ova pobolj\u0161anja su klju\u010dna za poljoprivrednike koji se oslanjaju na to\u010dne podatke za predvi\u0111anje prinosa i optimizaciju resursa. Na primjer, precizno brojanje klasova omogu\u0107uje bolje procjene proizvodnje \u017eitarica, informiraju\u0107i odluke o vremenu \u017eetve, skladi\u0161tenju i planiranju tr\u017ei\u0161ta.<\/p>\n

Budu\u0107i smjerovi i prakti\u010dne implikacije<\/h2>\n

Unato\u010d uspjehu, studija je priznala ograni\u010denja. Performanse su se smanjile u ekstremnim uvjetima osvjetljenja, poput o\u0161trog podnevnog odsjaja ili jakih sjena, \u0161to mo\u017ee zakloniti detalje \u0161iljaka. Osim toga, pravokutni okviri ponekad nisu odgovarali nepravilno oblikovanim \u0161iljcima, \u0161to je uzrokovalo manje neto\u010dnosti.<\/p>\n

Model je tako\u0111er isklju\u010dio mutne rubove iz slika bespilotnih letjelica, \u0161to je zahtijevalo ru\u010dnu predobradu - korak koji dodaje vrijeme i slo\u017eenost.<\/p>\n

Budu\u0107i rad ima za cilj rije\u0161iti ove probleme pro\u0161irivanjem skupa podataka kako bi uklju\u010dio slike snimljene u zoru, podne i sumrak, eksperimentiranjem s poligonalnim oznakama (fleksibilni oblici koji bolje odgovaraju nepravilnim objektima) i razvojem algoritama za bolje rukovanje mutnim podru\u010djima bez ru\u010dne intervencije.<\/p>\n

Implikacije ovog istra\u017eivanja su duboke. Za poljoprivrednike u regijama poput Tibeta, model nudi procjenu prinosa u stvarnom vremenu, zamjenjuju\u0107i radno intenzivno ru\u010dno brojanje automatizacijom temeljenom na dronovima. Razlikovanje faza rasta omogu\u0107uje precizno planiranje \u017eetve, smanjuju\u0107i gubitke od prerane ili odgo\u0111ene \u017eetve.<\/p>\n

Detaljni podaci o gusto\u0107i klasova - poput identificiranja nedovoljno naseljenih ili prenapu\u010denih podru\u010dja - mogu informirati strategije navodnjavanja i gnojidbe, smanjuju\u0107i otpad vode i kemikalija. Osim je\u010dma, lagana arhitektura obe\u0107ava i druge kulture, poput p\u0161enice, ri\u017ee ili vo\u0107a, otvaraju\u0107i put \u0161iroj primjeni u preciznoj poljoprivredi.<\/p>\n

Zaklju\u010dak<\/h2>\n

Zaklju\u010dno, ova studija ilustrira transformativni potencijal umjetne inteligencije u rje\u0161avanju poljoprivrednih izazova. Usavr\u0161avanjem YOLOv5 inovativnim laganim tehnikama, istra\u017eiva\u010di su stvorili alat koji uravnote\u017euje to\u010dnost i u\u010dinkovitost - \u0161to je klju\u010dno za primjenu u stvarnom svijetu u okru\u017eenjima s ograni\u010denim resursima.<\/p>\n

Pojmovi poput mAP-a, FLOP-ova i mehanizama pa\u017enje mogu se \u010diniti tehni\u010dkima, ali njihov utjecaj je duboko prakti\u010dan: omogu\u0107uju poljoprivrednicima dono\u0161enje odluka temeljenih na podacima, o\u010duvanje resursa i maksimiziranje prinosa. Kako klimatske promjene i rast stanovni\u0161tva poja\u010davaju pritisak na globalne prehrambene sustave, takav \u0107e napredak biti neophodan.<\/p>\n

Za poljoprivrednike Tibeta i \u0161ire, ova tehnologija ne predstavlja samo skok u poljoprivrednoj u\u010dinkovitosti, ve\u0107 i svjetionik nade za odr\u017eivu sigurnost hrane u neizvjesnoj budu\u0107nosti.<\/p>\n

Referenca: <\/strong>Cai, M., Deng, H., Cai, J. i dr. Detekcija laganog visokogorskog je\u010dma temeljena na pobolj\u0161anom YOLOv5. Plant Methods 21, 42 (2025). https:\/\/doi.org\/10.1186\/s13007-025-01353-0<\/a><\/p>","protected":false},"excerpt":{"rendered":"

Je\u010dam iz visokogorskih podru\u010dja, otporna \u017eitarica koja se uzgaja u visokogorskim regijama kineske visoravni Qinghai-Tibet, igra klju\u010dnu ulogu u lokalnoj sigurnosti hrane i gospodarstvu\u2026<\/p>","protected":false},"author":210249433,"featured_media":11562,"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":"","_crdt_document":"","content-type":"","_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","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":[1660,1657,1377],"tags":[],"class_list":["post-11559","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-agriculture-mapping","category-precision-farming","category-crop-monitoring"],"acf":[],"yoast_head":"\nBarley Farming Gets a Boost With Lightweight YOLOv5 Detection - 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\/hr\/blog\/uzgoj-jecma-dobiva-poticaj-laganom-detekcijom-yolov5\/\" \/>\n<meta property=\"og:locale\" content=\"hr_HR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection\" \/>\n<meta property=\"og:description\" content=\"Highland barley, a resilient cereal crop grown in the high-altitude regions of China\u2019s Qinghai-Tibet Plateau, plays a critical role in local food security and economic...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/geopard.tech\/hr\/blog\/uzgoj-jecma-dobiva-poticaj-laganom-detekcijom-yolov5\/\" \/>\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=\"2025-05-04T20:58:53+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-05-04T21:00:29+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection-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=\"Muhammad Farjad\" \/>\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=\"Napisao\/la\" \/>\n\t<meta name=\"twitter:data1\" content=\"Muhammad Farjad\" \/>\n\t<meta name=\"twitter:label2\" content=\"Procijenjeno vrijeme \u010ditanja\" \/>\n\t<meta name=\"twitter:data2\" content=\"8 minuta\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/\"},\"author\":{\"name\":\"Muhammad Farjad\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/person\\\/123c5562fb47aa8cf3aa81ae91e5e935\"},\"headline\":\"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection\",\"datePublished\":\"2025-05-04T20:58:53+00:00\",\"dateModified\":\"2025-05-04T21:00:29+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/\"},\"wordCount\":1570,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1\",\"articleSection\":[\"Agriculture Mapping\",\"Precision Farming\",\"Crop monitoring\"],\"inLanguage\":\"hr\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/\",\"url\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/\",\"name\":\"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection - GeoPard Agriculture\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1\",\"datePublished\":\"2025-05-04T20:58:53+00:00\",\"dateModified\":\"2025-05-04T21:00:29+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#breadcrumb\"},\"inLanguage\":\"hr\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"hr\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#primaryimage\",\"url\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1\",\"contentUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2025\\\/05\\\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1\",\"width\":3120,\"height\":1755,\"caption\":\"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/blog\\\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/geopard.tech\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection\"}]},{\"@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\":\"hr\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#organization\",\"name\":\"GeoPard Agriculture\",\"alternateName\":\"GeoPard\",\"url\":\"https:\\\/\\\/geopard.tech\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"hr\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2022\\\/03\\\/favicon.png?fit=200%2C200&ssl=1\",\"contentUrl\":\"https:\\\/\\\/i0.wp.com\\\/geopard.tech\\\/wp-content\\\/uploads\\\/2022\\\/03\\\/favicon.png?fit=200%2C200&ssl=1\",\"width\":200,\"height\":200,\"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\\\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/geopard.tech\\\/#\\\/schema\\\/person\\\/123c5562fb47aa8cf3aa81ae91e5e935\",\"name\":\"Muhammad Farjad\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"hr\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/51b9af87d9b0801ae1c0ab294a0f06eb669f60d5168f33394aeeca9de86537bb?s=96&d=identicon&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/51b9af87d9b0801ae1c0ab294a0f06eb669f60d5168f33394aeeca9de86537bb?s=96&d=identicon&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/51b9af87d9b0801ae1c0ab294a0f06eb669f60d5168f33394aeeca9de86537bb?s=96&d=identicon&r=g\",\"caption\":\"Muhammad Farjad\"},\"url\":\"#\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Uzgoj je\u010dma dobiva poticaj s laganom YOLOv5 detekcijom - 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\/hr\/blog\/uzgoj-jecma-dobiva-poticaj-laganom-detekcijom-yolov5\/","og_locale":"hr_HR","og_type":"article","og_title":"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection","og_description":"Highland barley, a resilient cereal crop grown in the high-altitude regions of China\u2019s Qinghai-Tibet Plateau, plays a critical role in local food security and economic...","og_url":"https:\/\/geopard.tech\/hr\/blog\/uzgoj-jecma-dobiva-poticaj-laganom-detekcijom-yolov5\/","og_site_name":"GeoPard - Precision agriculture Mapping software","article_publisher":"https:\/\/www.facebook.com\/geopardAgriculture\/","article_published_time":"2025-05-04T20:58:53+00:00","article_modified_time":"2025-05-04T21:00:29+00:00","og_image":[{"width":1024,"height":576,"url":"https:\/\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection-1024x576.png","type":"image\/png"}],"author":"Muhammad Farjad","twitter_card":"summary_large_image","twitter_creator":"@geopardagri","twitter_site":"@geopardagri","twitter_misc":{"Napisao\/la":"Muhammad Farjad","Procijenjeno vrijeme \u010ditanja":"8 minuta"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#article","isPartOf":{"@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/"},"author":{"name":"Muhammad Farjad","@id":"https:\/\/geopard.tech\/#\/schema\/person\/123c5562fb47aa8cf3aa81ae91e5e935"},"headline":"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection","datePublished":"2025-05-04T20:58:53+00:00","dateModified":"2025-05-04T21:00:29+00:00","mainEntityOfPage":{"@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/"},"wordCount":1570,"commentCount":0,"publisher":{"@id":"https:\/\/geopard.tech\/#organization"},"image":{"@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#primaryimage"},"thumbnailUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1","articleSection":["Agriculture Mapping","Precision Farming","Crop monitoring"],"inLanguage":"hr","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/","url":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/","name":"Uzgoj je\u010dma dobiva poticaj s laganom YOLOv5 detekcijom - GeoPard Agriculture","isPartOf":{"@id":"https:\/\/geopard.tech\/#website"},"primaryImageOfPage":{"@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#primaryimage"},"image":{"@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#primaryimage"},"thumbnailUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1","datePublished":"2025-05-04T20:58:53+00:00","dateModified":"2025-05-04T21:00:29+00:00","breadcrumb":{"@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#breadcrumb"},"inLanguage":"hr","potentialAction":[{"@type":"ReadAction","target":["https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/"]}]},{"@type":"ImageObject","inLanguage":"hr","@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#primaryimage","url":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1","contentUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1","width":3120,"height":1755,"caption":"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection"},{"@type":"BreadcrumbList","@id":"https:\/\/geopard.tech\/blog\/barley-farming-gets-a-boost-with-lightweight-yolov5-detection\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/geopard.tech\/"},{"@type":"ListItem","position":2,"name":"Barley Farming Gets a Boost With Lightweight YOLOv5 Detection"}]},{"@type":"WebSite","@id":"https:\/\/geopard.tech\/#website","url":"https:\/\/geopard.tech\/","name":"GeoPard - Softver za preciznu poljoprivredu","description":"Precizna poljoprivreda Softver za mapiranje","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":"hr"},{"@type":"Organization","@id":"https:\/\/geopard.tech\/#organization","name":"GeoPard Poljoprivreda","alternateName":"GeoPard","url":"https:\/\/geopard.tech\/","logo":{"@type":"ImageObject","inLanguage":"hr","@id":"https:\/\/geopard.tech\/#\/schema\/logo\/image\/","url":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2022\/03\/favicon.png?fit=200%2C200&ssl=1","contentUrl":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2022\/03\/favicon.png?fit=200%2C200&ssl=1","width":200,"height":200,"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\/"]},{"@type":"Person","@id":"https:\/\/geopard.tech\/#\/schema\/person\/123c5562fb47aa8cf3aa81ae91e5e935","name":"Muhamed Far\u017ead","image":{"@type":"ImageObject","inLanguage":"hr","@id":"https:\/\/secure.gravatar.com\/avatar\/51b9af87d9b0801ae1c0ab294a0f06eb669f60d5168f33394aeeca9de86537bb?s=96&d=identicon&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/51b9af87d9b0801ae1c0ab294a0f06eb669f60d5168f33394aeeca9de86537bb?s=96&d=identicon&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/51b9af87d9b0801ae1c0ab294a0f06eb669f60d5168f33394aeeca9de86537bb?s=96&d=identicon&r=g","caption":"Muhammad Farjad"},"url":"#"}]}},"jetpack_publicize_connections":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/geopard.tech\/wp-content\/uploads\/2025\/05\/Barley-Farming-Gets-a-Boost-With-Lightweight-YOLOv5-Detection.png?fit=3120%2C1755&ssl=1","jetpack_likes_enabled":true,"jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/pdiCPa-30r","_links":{"self":[{"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/posts\/11559","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/users\/210249433"}],"replies":[{"embeddable":true,"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/comments?post=11559"}],"version-history":[{"count":0,"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/posts\/11559\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/media\/11562"}],"wp:attachment":[{"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/media?parent=11559"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/categories?post=11559"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/geopard.tech\/hr\/wp-json\/wp\/v2\/tags?post=11559"}],"curies":[{"name":"radni list","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}