dr. sc. Ivana Rendulić Jelušić, Alti Agro d.o.o. , Zagreb

Doktorica znanosti iz područja vinogradarstva i vinarstva, certificirana enologinja i ovlaštena agronomkinja Hrvatske agronomske komore. Osnivačica i direktorica tvrtke ALTI AGRO d.o.o. te autorica nagrađivane AI platforme Vineyard Angel za precizno vinogradarstvo, a profesionalno djelovanje usmjerila je na povezivanje znanosti, tehnologije i prakse u razvoju rješenja za digitalnu transformaciju poljoprivrede. Vanjska je evaluatorica EU projekata i programa, s gotovo dvadeset godina iskustva u EU projektima, strateškom planiranju i razvoju inovacija

1.Prehrambena sigurnost

Kada govorimo o prehrambenoj sigurnosti, rasprava se najčešće vodi oko količine obradivog zemljišta, domaće proizvodnje, uvoza hrane ili dostupnosti vode i energije. Međutim, u uvjetima klimatskih promjena, sve skupljih inputa i kroničnog nedostatka radne snage, jednako važno pitanje postaje koliko učinkovito upravljamo proizvodnjom koju već imamo.

U takvim uvjetima više nije dovoljno samo proizvoditi. Potrebno je ranije prepoznati problem, preciznije odrediti gdje je intervencija stvarno potrebna i bolje iskoristiti resurse koji već stoje na raspolaganju. Na tom mjestu digitalizacija poljoprivrede iz tehnološkog trenda prelazi u pitanje konkurentnosti i otpornosti cijelog prehrambenog sustava.

Temeljna ideja precizne poljoprivrede je jednostavna: različiti dijelovi iste proizvodne površine nisu jednaki i zato njima ne bi uvijek trebalo upravljati na isti način. Unutar samo nekoliko hektara mogu postojati velike razlike u tlu, dostupnosti vode, bujnosti, zdravstvenom stanju i kvalitativnom potencijalu. Tradicionalni pristup te razlike pojednostavljuje: cijela se površina tretira istom količinom sredstva, prihranjuje jednakom dozom i bere prema jedinstvenom rasporedu.

Poljoprivrednik koji pravodobno zna gdje se unutar nasada pojavljuje stres, gdje nedostaju biljke, koji dijelovi zaostaju u razvoju ili gdje je povećan rizik od bolesti donosi bitno preciznije odluke od onoga koji cijelu površinu tretira istovjetno. To znači racionalnije korištenje vode, zaštitnih sredstava, gnojiva, energije i rada, ali i veću sposobnost očuvanja prinosa i kvalitete u sve nepredvidljivijim proizvodnim uvjetima.

2. Digitalna transformacija poljoprivrede

Digitalna transformacija poljoprivrede ne svodi se na nabavu opreme – drona, meteorološke postaje ili softvera. Vrijednost nastaje tek kada se prikupljeni podaci pretvore u informaciju na temelju koje agronom ili proizvođač može donijeti bolju odluku. Takav pristup danas postoji i u Hrvatskoj. Tvrtka Alti Agro u suradnji s TIS Grupom razvila je Vineyard Angel, digitalnu platformu za precizno upravljanje vinogradarskom proizvodnjom, koja multispektralne podatke i umjetnu inteligenciju koristi za analizu stanja vinove loze tijekom vegetacijske sezone.

Vinogradarstvo pritom dobro pokazuje zašto je važno kako se podaci prikupljaju i obrađuju. Vinova loza raste u redovima i, gledano iz zraka, zauzima razmjerno mali dio nasada, dok znatno veću površinu čine među redni prostor, trava, korovi i tlo. Ako se snimka analizira kao jedinstvena vegetacijska masa – što je čest pristup – stanje vinove loze miješa se s pratećom vegetacijom, a proizvođač dobiva pogrešnu informaciju o nasadu. Vineyard Angel zato uz pomoć umjetne inteligencije iz multispektralne snimke izdvaja isključivo redove vinove loze i vegetacijske indekse računa samo na vegetaciji od interesa, bez utjecaja među redne vegetacije. Rezultat je detaljnija i pouzdanija slika stvarnog stanja nasada. Na istoj se platformi prate razlike u bujnosti, provodi kvalitativno zoniranje, geolociraju nedostajući trsovi te koriste meteorološki podaci i prediktivni modeli za najvažnije bolesti.

Sljedeći je korak povezivanje analize s intervencijom u nasadu. Umjesto da cijeli nasad primi istu količinu sredstva, digitalna karta zona postaje osnova za varijabilnu primjenu, pri kojoj svaka zona dobiva količinu koja odgovara njezinom stvarnom stanju. U vinogradarstvu se to razvija u smjeru varijabilnih karata za tretiranje koje mogu koristiti i dron i traktorska prskalica. Time se zatvara krug precizne poljoprivrede: mjerenje – analiza – stručna odluka – ciljano djelovanje.

Dronovi su pritom praktičan alat. Mogu brzo snimiti velike površine i prikupiti podatke za analizu, a mogu se koristiti i za pojedine zahvate poput folijarne prihrane ili tretiranja kada je takva primjena agronomski opravdana. Njihova je vrijednost u tome što su dio dobro organiziranog procesa – od prikupljanja podataka do ciljane intervencije u nasadu.

Isto vrijedi i za umjetnu inteligenciju. Algoritam prepoznaje uzorke, izdvaja anomalije i brzo obrađuje veliku količinu podataka, ali vinograd ne poznaje na način na koji ga poznaje čovjek koji njime upravlja. Ne poznaje povijest nasada, proizvodni cilj, specifičnosti sorte i položaja ni sve okolnosti koje ulaze u odluku. Konačna procjena zato ostaje stručna odluka agronoma, utemeljena na pravilno prikupljenim i obrađenim podacima.

3. Učinkovitost korištenja resursa i energija

U kontekstu prehrambene sigurnosti posebno je važna učinkovitost korištenja resursa, a ona je izravno povezana s energijom. Poljoprivreda ovisi o energiji ne samo kroz gorivo za mehanizaciju i električnu energiju za navodnjavanje i hlađenje, nego i posredno – kroz mineralna gnojiva, sredstva za zaštitu bilja i logistički lanac koji stoji iza njih. Svaki poremećaj u dostupnosti ili cijeni jednog od tih inputa prenosi se na proizvodnju hrane. Smanjivanje nepotrebne potrošnje zato odavno više nije samo ekološko pitanje, nego pitanje ekonomske otpornosti proizvodnje i energetske izloženosti prehrambenog sektora.

Iskustva iz prakse i istraživanja u preciznom vinogradarstvu pokazuju da varijabilno tretiranje može smanjiti potrošnju sredstava za zaštitu bilja i folijarnih gnojiva za 15 do 30 %, uz razmjerno manju potrošnju vode za pripremu škropiva. Kada se tome doda pravodobnost koja proizlazi iz prediktivnih modela – tretiranje onda kada je stvarno potrebno, umjesto po kalendaru – ušteda dodatno raste. Ušteda inputa pritom nije ni jedina ni nužno najveća korist, iako se najlakše izračuna. Bolje poznavanje varijabilnosti unutar nasada smanjuje rizik od pogrešne odluke, a kvalitativno zoniranje omogućuje selektivnu berbu kojom se iz istog vinograda mogu dobiti dvije ili više kvalitativnih kategorija grožđa. Te se koristi rjeđe kvantificiraju jer se ne pojavljuju kao niži trošak, nego kao veća vrijednost proizvoda i manji broj pogrešaka.

4.Odluke u pravom trenutku

U poljoprivredi često nije dovoljno donijeti ispravnu odluku – treba je donijeti i provesti u pravom trenutku. Kod zaštite bilja, navodnjavanja, prihrane ili berbe optimalno vrijeme za intervenciju mjeri se danima, a ponekad i satima. Snimanje površine od 20 do 30 hektara traje svega pola sata, a rezultati analize dostupni su unutar nekoliko sati. Klasičan obilazak i vizualna procjena istog nasada oduzeli bi nekoliko dana rada, a rezultat bi bila procjena, dok se snimanjem varijabilnost stvarno mjeri i kvantificira. Agronomu to ne oduzima posao, nego mu omogućuje da vrijeme usmjeri upravo na dijelove nasada u kojima postoji problem.

5.Digitalizacija u odnosu na suverenost

Digitalizacija ima i dimenziju koja se tiče suverenosti. Ako će podaci, senzori, algoritmi i digitalne platforme imati sve veću ulogu u upravljanju proizvodnjom hrane, onda nije svejedno tko te tehnologije razvija, gdje se nalaze podaci i koliko su rješenja prilagođena stvarnim potrebama domaćih proizvođača.

Vineyard Angel danas se koristi na više od 800 hektara vinograda u Hrvatskoj i regiji, kod velikih proizvođača, ali i manjih vinarija usmjerenih na vina visoke kvalitete. To pokazuje da razvoj poljoprivredne tehnologije u Hrvatskoj ne mora ostati na razini istraživačkog ili pilot-projekta, nego može postati stvaran proizvodni alat. Razvoj takvih rješenja usto povezuje sektore koji su se donedavno promatrali odvojeno: poljoprivredu, informatiku, umjetnu inteligenciju, meteorologiju, strojarstvo i energetiku. Ta će interdisciplinarnost biti jedna od glavnih karakteristika poljoprivrede budućnosti.

6.Što danas nije pitanje?

Pitanje danas nije treba li digitalizirati poljoprivredu, nego koliko brzo možemo prijeći s izoliranih tehnoloških rješenja na sustave koji proizvođaču omogućuju preciznije upravljanje resursima i proizvodnim rizicima. Klimatske promjene neće poljoprivredu učiniti jednostavnijom. Nedostatak radne snage neće nestati. Voda, energija, gnojiva i sredstva za zaštitu bilja neće postati manje važni. Zbog toga će sposobnost da iz svakog hektara, svakog kubičnog metra vode, svakog kilograma inputa i svakog sata rada dobijemo veću proizvodnu vrijednost postati jedan od temelja konkurentnosti europske i hrvatske poljoprivrede.

Prehrambena sigurnost ne počinje tek u trenutku kada hrane nedostaje. Gradi se godinama ranije – ulaganjem u znanje, tehnologiju i domaću proizvodnju.

Dronovi, senzori i umjetna inteligencija pritom nisu cilj sami po sebi. Oni su alati za stvaranje otpornije poljoprivrede, a otpornija poljoprivreda jedan je od preduvjeta stvarne prehrambene suverenosti.

Dr. Ivana Rendulić Jelušić, Alti Agro doo, Zagreb

Doctor of Science in the field of viticulture and winemaking, certified enologist and authorized agronomist of the Croatian Chamber of Agriculture. Founder and CEO of ALTI AGRO doo and author of the award-winning AI platform Vineyard Angel for precision viticulture, and her professional work is focused on connecting science, technology and practice in developing solutions for the digital transformation of agriculture. She is an external evaluator of EU projects and programs, with almost twenty years of experience in EU projects, strategic planning and innovation development

 
Digitalization of agriculture as a new pillar of food security

1.Food safety

When we talk about food security, the debate usually revolves around the amount of arable land, domestic production, food imports or the availability of water and energy. However, in the face of climate change, increasingly expensive inputs and chronic labor shortages, an equally important question is how effectively we manage the production we already have.

In such conditions, it is no longer enough to just produce. It is necessary to recognize the problem earlier, determine more precisely where intervention is really needed, and make better use of the resources already available. At this point, the digitalization of agriculture moves from a technological trend to a question of the competitiveness and resilience of the entire food system.

The basic idea behind precision agriculture is simple: different parts of the same production area are not the same and therefore should not always be managed in the same way. Within just a few hectares, there can be huge differences in soil, water availability, vigor, health status and quality potential. The traditional approach simplifies these differences: the entire area is treated with the same amount of product, fertilized at the same rate and harvested according to a single schedule.

A farmer who knows in a timely manner where stress occurs within the plantation, where plants are missing, which parts are lagging behind in development or where the risk of disease is increased makes significantly more precise decisions than one who treats the entire area identically. This means more rational use of water, protective agents, fertilizers, energy and labor, but also a greater ability to preserve yield and quality in increasingly unpredictable production conditions.

2. Digital transformation of agriculture

The digital transformation of agriculture is not limited to the purchase of equipment – a drone, a weather station or software. Value is created only when the collected data is converted into information on the basis of which the agronomist or producer can make a better decision. Such an approach also exists in Croatia today. The company Alti Agro, in cooperation with the TIS Group, has developed Vineyard Angel , a digital platform for precise management of vineyard production, which uses multispectral data and artificial intelligence to analyze the condition of the vine during the growing season.

Viticulture shows well why it is important how data is collected and processed. The vines grow in rows and, seen from the air, occupy a relatively small part of the plantation, while a much larger area is made up of inter-row space, grass, weeds and soil. If the recording is analyzed as a single mass of vegetation – which is a common approach – the condition of the vines is mixed with the accompanying vegetation, and the producer gets the wrong information about the plantation. That’s why Vineyard Angel, with the help of artificial intelligence, extracts only the rows of vines from the multispectral image and calculates vegetation indices only on the vegetation of interest, without affecting the rows of vegetation. The result is a more detailed and reliable picture of the real state of the plantations. On the same platform, differences in lushness are monitored, qualitative zoning is carried out, missing vines are geolocated, and meteorological data and predictive models for the most important diseases are used.

The next step is to link the analysis to the intervention in the plantation. Instead of the entire plantation receiving the same amount of product, the digital zone map becomes the basis for variable application, in which each zone receives the amount that corresponds to its actual condition. In viticulture, this is developing in the direction of variable treatment maps that can be used by both drones and tractor sprayers. This closes the circle of precision agriculture: measurement – analysis – expert decision – targeted action.

Drones are a practical tool. They can quickly capture large areas and collect data for analysis, and can also be used for individual interventions such as foliar fertilization or treatments when such application is agronomically justified. Their value lies in being part of a well-organized process – from data collection to targeted intervention in the plantation.

The same goes for artificial intelligence. The algorithm recognizes patterns, extracts anomalies, and quickly processes large amounts of data, but it does not know the vineyard in the way that the human who manages it knows it. It does not know the history of the plantations, the production goal, the specifics of the variety and location, or all the circumstances that go into the decision. The final assessment therefore remains the professional decision of the agronomist, based on properly collected and processed data.

3. Resource and energy efficiency

In the context of food security, resource efficiency is particularly important , and it is directly linked to energy. Agriculture depends on energy not only through fuel for machinery and electricity for irrigation and cooling, but also indirectly – through mineral fertilizers, plant protection products and the logistics chain that stands behind them. Any disruption in the availability or price of one of these inputs is transmitted to food production. Reducing unnecessary consumption is therefore no longer just an ecological issue, but a question of the economic resilience of production and the energy exposure of the food sector.

Experience from practice and research in precision viticulture shows that variable treatment can reduce the consumption of plant protection products and foliar fertilizers by 15 to 30%, with a relatively lower consumption of water for preparing the spray. When the timeliness resulting from predictive models is added to this – treatment when it is actually needed, instead of according to the calendar – the savings increase further. Input savings are neither the only nor necessarily the greatest benefit, although they are the easiest to calculate. Better knowledge of variability within the plantation reduces the risk of making the wrong decision, and qualitative zoning allows for selective harvesting that can obtain two or more quality categories of grapes from the same vineyard. These benefits are less frequently quantified because they do not appear as lower costs, but as higher product value and fewer errors.

4. Decisions at the right time

In agriculture, it is often not enough to make the right decision – it must be made and implemented at the right time. When it comes to plant protection, irrigation, feeding or harvesting, the optimal time for intervention is measured in days and sometimes hours. Recording an area of 20 to 30 hectares takes only half an hour, and the results of the analysis are available within a few hours. A classic tour and a visual assessment of the same plantation would take several days of work, and the result would be an assessment, while with recording the variability is actually measured and quantified. This does not take away the work of the agronomist, but allows him to focus his time precisely on the parts of the plantation where there is a problem.

5.Digitalization in relation to sovereignty

Digitization also has a dimension that concerns sovereignty. If data, sensors, algorithms and digital platforms will play an increasing role in the management of food production, then it does not matter who develops these technologies, where the data is located and how well the solutions are adapted to the real needs of domestic producers.

Vineyard Angel is currently used on more than 800 hectares of vineyards in Croatia and the region, by large producers, as well as smaller wineries focused on high-quality wines. This shows that the development of agricultural technology in Croatia does not have to remain at the level of research or pilot projects, but can become a real production tool. The development of such solutions also connects sectors that were previously considered separate: agriculture, informatics, artificial intelligence, meteorology, mechanical engineering and energy. This interdisciplinarity will be one of the main characteristics of the agriculture of the future.

6.What is not a question today?

The question today is not whether to digitize agriculture , but how quickly we can move from isolated technological solutions to systems that enable producers to manage resources and production risks more precisely. Climate change will not make agriculture easier. Labor shortages will not disappear. Water, energy, fertilizers and plant protection products will not become less important. Therefore, the ability to extract greater production value from every hectare, every cubic meter of water, every kilogram of input and every hour of work will become one of the foundations of the competitiveness of European and Croatian agriculture.

Food security does not start until the moment when food is scarce. It is being built years earlier – by investing in knowledge, technology and domestic production.

Drones, sensors and artificial intelligence are not a goal in themselves. They are tools for creating more resilient agriculture, and more resilient agriculture is one of the prerequisites for real food sovereignty.