Routes to dig­i­tal­i­sa­tion in agriculture

Over the past few decades, agri­cul­ture has evolved from being a sec­tor that was once heav­i­ly reliant on man­u­al labour to one that is now high­ly mech­a­nised. Mech­a­ni­sa­tion leads to a rise in com­plex­i­ty, mean­ing more and more deci­sions have to be made simul­ta­ne­ous­ly. Work­ing with pre­ci­sion requires sophis­ti­cat­ed sen­sors and deci­sion-mak­ing tools, and sys­tems that are capa­ble of inter-com­mu­ni­ca­tion. To main­tain vis­i­bil­i­ty and con­trol, process­es are large­ly auto­mat­ed and work steps are doc­u­ment­ed to a high degree of accu­ra­cy. Fields are extreme­ly vari­able in terms of their regrowth and soil struc­ture, which means a great deal of deci­sions need to be made in order to farm them effi­cient­ly. This is exact­ly where the ben­e­fits of pre­ci­sion farm­ing or pre­ci­sion agri­cul­ture come in. The pro­po­nents of this devel­op­ment are politi­cians, sci­en­tists, soci­ety as a whole and the agri­cul­tur­al indus­try itself. Con­tex­tu­al sci­en­tif­ic knowl­edge of crop grow­ing indi­cates areas where resources can be saved and green­house gas­es reduced, while polit­i­cal frame­works and soci­etal demands deter­mine the play­ing field and set the course. The grow­ing soci­etal demands and legal require­ments in farm­ing have the poten­tial to be solved through rapid­ly evolv­ing tech­nol­o­gy. The ris­ing price pres­sure as a result of inter­na­tion­al com­pe­ti­tion also requires an indus­try-wide boost in pro­duc­tion effi­cien­cy for agri­cul­tur­al goods.

Dri­ving forces for change 

The push for dig­i­tal­i­sa­tion has pen­e­trat­ed all areas of agri­cul­ture, so the indus­try is not alone in fac­ing its chal­lenges. Agri­cul­ture also ben­e­fits from the progress that is being made, for exam­ple in infor­ma­tion tech­nol­o­gy. Posi­tion­ing sys­tems for deter­min­ing exact loca­tions are a core exam­ple of progress in this area. Freely avail­able satel­lite data has enabled the tech­nol­o­gy to come on in leaps and bounds. The loca­tion data of machines can be accu­rate­ly mea­sured to a few cen­time­tres through a com­bi­na­tion of var­i­ous process­es. Tech­ni­cal solu­tions for elec­tron­ic machine con­trols, such as part width sec­tion con­trol in plant pro­tec­tion, as well as dri­ve con­trols and meter­ing equip­ment in fer­tilis­er spread­ers and seed drills, can all ben­e­fit from this. Spe­cial sen­sors are also required to record the sta­tus of crops, for exam­ple. The dig­i­tal­i­sa­tion path in arable farm­ing leads to a process chain of data-dri­ven crop production.

The mar­ket for pre­ci­sion agri­cul­ture is now wide­ly devel­oped and there are var­i­ous providers of com­plete farm man­age­ment infor­ma­tion sys­tems, which are not only use­ful on the field, but also in the barn and across the entire farm. So much so that it begs the ques­tion – where do I start?! And what are the issues and chal­lenges with­in this domain?

The first chal­lenge usu­al­ly begins dur­ing imple­men­ta­tion. Con­vert­ing to a new sys­tem involves invest­ing a great deal of time in the begin­ning. Then there are the finan­cial risks. This is where a cost-ben­e­fit analy­sis must be car­ried out. The deci­sion on which appli­ca­tions to use and whether they will be prac­ti­ca­ble in the long term is anoth­er big chal­lenge. Sim­ply gath­er­ing data is not enough on its own; the data also needs to be visu­alised and eval­u­at­ed. Depend­ing on the infor­ma­tion pro­duced, activ­i­ties can then be planned and imple­ment­ed as a next step. This requires suit­able com­mu­ni­ca­tion chan­nels, inter­faces, machines and imple­ments. After all, an appli­ca­tion map isn’t much use to an ox. In essence, farms need to be aware that there are mul­ti­ple fac­tors and var­i­ous sys­tems at play when it comes to implementation.

The sec­ond biggest chal­lenge comes in adapt­ing the sys­tem to local con­di­tions. Net­work cov­er­age is often prob­lem­at­ic. In terms of data record­ing, the func­tions should ide­al­ly be set up in a way that enables data to be saved to each mobile device then syn­chro­nised with the asso­ci­at­ed cloud stor­age sys­tem once they are back online. Local con­di­tions also incor­po­rate the spe­cif­ic farm struc­ture. The con­nec­tiv­i­ty of sys­tems and machines is often a hur­dle in this case. Agri­cul­tur­al con­trac­tors, in par­tic­u­lar, have mixed fleets and some of their machines are more dig­i­tal than oth­ers. Coor­di­nat­ing the indi­vid­ual com­po­nents is cru­cial and presents a huge chal­lenge. If one machine is already high­ly dig­i­talised, the sur­round­ing machines often have to catch up to a cer­tain degree in order to be able to com­mu­ni­cate smooth­ly. Stan­dard­i­s­a­tions, man­u­fac­tur­er-inde­pen­dent appli­ca­tions and options for con­nect­ing ana­logue farm machin­ery offer one solu­tion to the prob­lem. Those oper­at­ing the machines must also be tak­en into con­sid­er­a­tion. They need intu­itive pro­grammes that offer clear, user-friend­ly functions.

Cur­rent areas of appli­ca­tion and process­es in agriculture

There is no exact def­i­n­i­tion of pre­ci­sion agri­cul­ture or pre­ci­sion farm­ing that is recog­nised across the board. The terms are used to describe var­i­ous areas of agri­cul­ture. The con­cept rep­re­sents more than just tech­nol­o­gy for deal­ing with vari­able local con­di­tions – it is also about design­ing the entire process chain and man­ag­ing infor­ma­tion. It involves a com­bi­na­tion of dif­fer­ent com­po­nents. Pre­ci­sion farm­ing incor­po­rates auto­mat­ic data record­ing as a basis, sec­tion-spe­cif­ic equip­ment for all crop cul­ti­va­tion activ­i­ties, fleet man­age­ment (loca­tion, route plan­ning, machine data mon­i­tor­ing) and field robot­ics (auto­mat­ed imple­ment con­trol for manned and unmanned vehicles).

In prin­ci­ple, a dis­tinc­tion is made between online and offline process­es. Field poten­tial and appli­ca­tion maps are gen­er­at­ed in advance then trans­mit­ted – often man­u­al­ly – in sep­a­rate work steps to the trac­tor’s ter­mi­nal. This is what is referred to as the offline process. The machines are then guid­ed by the stored GPS data to per­form the crop cul­ti­va­tion tasks. There­fore, hav­ing a sta­ble GPS sig­nal is essen­tial for autonomous dri­ving and steered imple­ments. As part of the online process, the data is cap­tured at the exact moment the task is car­ried out. Dur­ing plant pro­tec­tion activ­i­ties, for exam­ple, infor­ma­tion about the crop is record­ed by sen­sor equip­ment mount­ed on the front of the trac­tor. Based on the infor­ma­tion gath­ered, the con­trols for the attached sprayer are influ­enced via spe­cif­ic tar­get val­ues. The infor­ma­tion is exchanged in a split sec­ond via elec­tron­ic com­mu­ni­ca­tions between the rel­e­vant com­po­nents. Sim­i­lar process­es can be used when spread­ing fer­tilis­er. Good sen­sor equip­ment that draws on dif­fer­ent process­es is there­fore required. This enables infor­ma­tion on the nitro­gen sup­ply, weed infes­ta­tion, bio­mass lev­el and dis­ease infes­ta­tion to be detect­ed. Opti­cal sen­sors utilise the prop­er­ty of plants to absorb and reflect less red light dur­ing pho­to­syn­the­sis. This enables the active bio­mass to be deter­mined. In near infrared spec­troscopy (NIR), the oppo­site effect is used. Plant struc­tures reflect near infrared light, which is not present on the spec­trum vis­i­ble to human eyes. It pro­duces a char­ac­ter­is­tic reflec­tion spec­trum, which indi­cates the con­di­tion of the crops on the field. To assess the pres­ence of dis­ease, ther­mal sen­sors can be used to mea­sure the sur­face tem­per­a­ture of leaves, which enables healthy leaves to be dis­tin­guished from dis­eased leaves when a fun­gal infes­ta­tion is present.

In the past, farm­ers have often been accused of not pro­duc­ing enough doc­u­men­ta­tion. A lack of time, valid mea­sure­ments and stan­dard­ised for­mats has tend­ed to be the rea­son. Ulti­mate­ly, the data also needs to be eval­u­at­ed and inter­pret­ed. Things are much sim­pler now thanks to the mod­ern tech­nol­o­gy. In order to trans­form the mea­sured val­ues into infor­ma­tion, the data must first be record­ed and stored. In pre­ci­sion agri­cul­ture, this takes place auto­mat­i­cal­ly as each task is com­plet­ed. Using the machine’s GPS sig­nal, the data can also be allo­cat­ed to a spe­cif­ic posi­tion on the field and pre­sent­ed on var­i­ous maps. The main areas of appli­ca­tion are yield maps, which are based on infor­ma­tion gath­ered from har­vesters, and veg­e­ta­tion maps, which incor­po­rate satel­lite data. Anoth­er option is to incor­po­rate soil sam­pling results so they can be illus­trat­ed on the field map. These map resources enable crop cul­ti­va­tion activ­i­ties to be planned and imple­ment­ed in a tar­get­ed manner.

In the future, the momen­tum for dig­i­tal­i­sa­tion is bound to be even stronger. The devel­op­ment of field and barn robots on farms is far from over. Robots are often seen in barns nowa­days, but they’re less com­mon in the field. Machines are capa­ble of sow­ing a field or con­trol­ling weeds with the help of sen­sors and achieve a sim­i­lar acreage per­for­mance to their larg­er rel­a­tives. There is still plen­ty of room for devel­op­ment in the robot­ics domain. In the soft­ware domain, providers who can adapt dynam­i­cal­ly to the needs of both their users and the indus­try are the ones that will stay the course.