Twenty years of Aus­tralasian pre­ci­sion agri­cul­ture _______________________________

“The pur­pose of pre­ci­sion agri­cul­ture has al­ways been to in­crease the num­ber of cor­rect de­ci­sions made in the busi­nesses of crop and an­i­mal man­age­ment,” Brett Whe­lan told del­e­gates at the 20th Sym­po­sium on Pre­ci­sion Agri­cul­ture (PA) in Sydney.

NZ Grower - - Election 2017 - Dan Bloomer

“It is a log­i­cal step in the evo­lu­tion of agri­cul­tural man­age­ment sys­tems to­ward in­creased ef­fi­ciency of in­puts rel­a­tive to pro­duc­tion, min­imised waste and im­proved prod­uct qual­ity, trace­abil­ity and mar­ketabil­ity,” he said.

The PA Sym­po­sium brings to­gether farm­ers, grow­ers, re­searchers, ad­vi­sors and in­dus­try to dis­cuss and ab­sorb de­vel­op­ments. Speak­ers cov­ered cut­ting edge re­search, on­farm ap­pli­ca­tion by re­searchers, ad­vi­sors and farm­ers, and in­dus­try back­ground in­for­ma­tion such as the state of telecom­mu­ni­ca­tions and data own­er­ship.

Crop and soil sens­ing con­tin­ues to de­velop, and there is in­creas­ing use of new ap­proaches. Canopy as­sess­ment has re­lied heav­ily on NDVI (Nor­malised Dif­fer­ence Veg­e­ta­tion In­dex), the 1970s veg­e­ta­tion in­dex cho­sen for dis­tin­guish­ing for­est from desert and ocean. In re­cent years a wider range of sen­sors cap­tur­ing more light bands (blue, green, red and in­frared) have be­come af­ford­able and avail­able. Some look at red-edge and ther­mal infra-red, two bands of­ten re­lated to crop stress of some form. Off the shelf cam­eras that fit sim­ple UAVs (Un­manned Aerial Ve­hi­cles) are within farm bud­gets now.

Ian Yule de­scribed re­search with hy­per­spec­tral sen­sors that cap­ture very de­tailed im­ages with hun­dreds of light bands. Hun­dreds of ground con­trol sam­ples pro­vide “real” in­for­ma­tion and enor­mous amounts of data get an­a­lysed to iden­tify re­la­tion­ships. The ca­pac­ity of this to de­ter­mine species,

Speak­ers cov­ered cut­ting edge re­search, on­farm ap­pli­ca­tion by re­searchers, ad­vi­sors and farm­ers, and in­dus­try back­ground in­for­ma­tion such as the state of telecom­mu­ni­ca­tions and data own­er­ship.

plant nu­tri­ent sta­tus and other use­ful in­for­ma­tion is re­mark­able. The cur­rent re­search equip­ment and pro­cess­ing is very ex­pen­sive but as­sume price drops as com­mer­cial­i­sa­tion pro­gresses.

Ma­chine vi­sion in­clud­ing object shape, tex­ture and colour is be­ing used to recog­nise in­di­vid­ual ob­jects such as plants, parts of plants or spe­cific weeds. Dis­cussing robotics re­search to guide de­ci­sion mak­ing on veg­etable farms, Zhe Xu noted: “If a hu­man can recog­nise some­thing, a ma­chine can be taught to as well.” Get used to ar­ti­fi­cial in­tel­li­gence, neu­ral pro­gram­ming and au­ton­o­mous phe­no­typ­ing!

We pre­sented our own onions re­search which is us­ing smart­phone cam­eras to cap­ture very use­ful crop de­vel­op­ment in­for­ma­tion quickly and cost ef­fec­tively. Com­bined with crop mod­els and web based cal­cu­la­tion we can pre­dict final yields with fair ac­cu­racy early enough to sup­port crop man­age­ment de­ci­sions.

An Aus­tralian veg­etable re­search project is us­ing sim­i­lar ap­proaches to sup­port de­ci­sion mak­ing in car­rot crops and in­ves­ti­gat­ing oth­ers with prom­ise. That team in­cludes re­searchers and

farm­ers, and is in­creas­ingly us­ing yield mon­i­tors for crops such as pota­toes and car­rots. Con­vert­ing yield data to value al­lows farm­ers to es­ti­mate costs of vari­abil­ity and how much to invest to fix prob­lem ar­eas.

Data cap­ture, com­mu­ni­ca­tions and analysis was a key theme. Kim Bryce­son de­scribed the es­tab­lish­ment of a sen­sor net­work and an­a­lyt­ics us­ing IoT (in­ter­net of things) tools at Queens­land Univer­sity Gat­ton. Rob Bram­ley ex­plained a process that pre­dicted sugar yields at re­gional scale to pro­mote bet­ter fer­tiliser man­age­ment in that in­dus­try. Pa­trick Filippi pre­sented a “big data” ap­proach to pre­dict­ing grain yield. The data rev­o­lu­tion is chang­ing our world in ways we can’t yet imag­ine. The in­creas­ing amount of things mea­sured, the spa­tial scale and time span of col­lec­tion and de­vel­op­ment of data sci­ence to an­a­lyse huge streams of in­for­ma­tion rev­o­lu­tionise our un­der­stand­ing. These are ex­cit­ing times. Some jobs are go­ing to go, but oth­ers will be cre­ated as we re­quire com­pletely new skills for jobs not heard of a decade ago.

“We are all in the po­si­tion of mak­ing de­ci­sions from a limited un­der­stand­ing or a par­tic­u­lar per­spec­tive, work­ing with bi­o­log­i­cal sys­tems that are in­cred­i­bly com­plex and im­pos­si­ble to fully un­der­stand, “said Ian Yule. “Re­cent ex­pe­ri­ence with new sens­ing tech­nolo­gies and data pro­cess­ing has pro­duced new in­for­ma­tion that chal­lenges our pre­con­ceived ideas and un­der­stand­ings,” he said.

The PA Sym­po­sium is pre­sented by SPAA, the So­ci­ety for Pre­ci­sion Agri­cul­ture Aus­tralia, and the Pre­ci­sion Agri­cul­ture Lab­o­ra­tory at the Univer­sity of Sydney. There has al­ways been a New Zealand pres­ence be­cause while some de­tails are unique, the tools and pro­cesses are for the most part generic.

▴ The stan­dard RGB cam­era (white) at rear is sup­ple­mented by a Par­rot Se­quoia crop health analysis multi­band cam­era (black) in front. Many of such units use mul­ti­ple cam­eras with a se­lec­tion of wave­band fil­ters al­low­ing users to choose nu­mer­ous crop in­dices de­pend­ing on pur­pose. ◀ Con­sumer UAVs with off the shelf sen­sors are avail­able at a price farms can af­ford. Pro­cess­ing tech­nol­ogy is read­ily avail­able and within ca­pa­bil­ity of com­puter savvy op­er­a­tors. This unit has a sen­sor on top to man­age chang­ing light con­di­tions.

Newspapers in English

Newspapers from New Zealand

© PressReader. All rights reserved.