[{"data":1,"prerenderedAt":188},["ShallowReactive",2],{"en-post-dji-onboard-ai-challenge-2026-winners":3,"en-related-dji-onboard-ai-challenge-2026-winners":160},{"id":4,"title":5,"author":6,"body":7,"category":140,"date":141,"description":142,"draft":143,"extension":144,"image":145,"meta":146,"navigation":147,"path":148,"seo":149,"slug":150,"stem":151,"tags":152,"translation_slug":133,"updated":158,"__hash__":159},"en_posts\u002Fen\u002Fposts\u002Fdji-onboard-ai-challenge-2026-winners.md","DJI Picks 15 Winners in Global Onboard AI Challenge","Lucas Buzzo",{"type":8,"value":9,"toc":132},"minimark",[10,14,17,22,25,33,42,44,48,51,54,74,77,79,83,91,94,96,100,103,106,108,112,114],[11,12,13],"p",{},"DJI announced the winners of its Enterprise Drone Onboard AI Challenge 2026 on August 20, naming 5 Best Onboard AI Model Award recipients and 10 Industry Application Excellence Award recipients out of a global field of developers. Every winning project runs its AI model directly on the drone's onboard computer instead of streaming footage to the cloud for processing.",[15,16],"hr",{},[18,19,21],"h2",{"id":20},"background-why-dji-launched-an-onboard-ai-challenge","Background: Why DJI Launched an Onboard AI Challenge",[11,23,24],{},"DJI opened the contest in March 2026 to push developers beyond using drones as flying cameras that simply collect data for someone else to analyze later. The challenge instead asked teams to build AI models that run on the aircraft itself, turning raw video into a finished decision — a crack flagged, a crop counted, a person located — before the drone even lands.",[11,26,27,28,32],{},"The distinction matters technically. Cloud-based AI requires a live data link back to a server, which adds latency and fails wherever connectivity is weak — a common condition on farms, bridges, and disaster sites. ",[29,30,31],"strong",{},"Onboard AI"," is processing that happens on hardware carried by the drone itself, using no external network connection. DJI opened its software development kit (SDK) and the onboard compute of its enterprise aircraft to contestants, who submitted solutions through May 31, 2026, ahead of the August judging round.",[11,34,35,36,41],{},"Ten judges from the AI and drone industries scored submissions on technical sophistication and real-world deployment potential. Eligible hardware included the Matrice 4 and 4D series, Dock 3, Matrice 400, and DJI's Manifold 3 onboard computer — the same enterprise line covered in our ",[37,38,40],"a",{"href":39},"\u002Fdji-drones-complete-guide\u002F","complete guide to DJI's drone lineup",".",[15,43],{},[18,45,47],{"id":46},"who-won-15-projects-from-crop-counting-to-search-and-rescue","Who Won: 15 Projects From Crop Counting to Search and Rescue",[11,49,50],{},"The five Best Onboard AI Model Award winners were Aoshi Cangqiong, Cheng'an Zhilian (Chongqing) Information Technology, Hangzhou New Modal Technology, Huineng Technology, and Zhongke Beiwei (Beijing) Technology. The ten Industry Application Excellence Award winners spanned agriculture, infrastructure inspection, environmental monitoring, and emergency response, including Bridge Inspection Master, Chongqing Guangruida Technology, independent developer Daniel Tovar, East China Normal University's Estuarine and Coastal Data Center team, Lanzhou University's Grass Flying Team, Muniu Hetu Algorithm Team, Shangtejie's AI R&D department, the Shenzhen Public Welfare Rescue Volunteer Federation, Sky Patrol Smart Inspection, and Wuhan Luanchuang Digital Technology.",[11,52,53],{},"Two projects illustrate what \"onboard\" buys in practice:",[55,56,57,68],"ul",{},[58,59,60,63,64,41],"li",{},[29,61,62],{},"AgroCount AI",", built by Colombia-based developer Daniel Tovar, counts banana plants across a 50-hectare plantation using onboard processing — work DJI says previously took four to six field workers as long as four days, a task relevant to operators weighing drones for ",[37,65,67],{"href":66},"\u002Fagricultural-drone-guide\u002F","agricultural surveying",[58,69,70,73],{},[29,71,72],{},"The 9-in-1 Onboard Fusion Algorithm",", from Hangzhou New Modal Technology, bundles nine separate inspection functions — covering traffic monitoring, road maintenance, and enforcement — into a single model that feeds results straight into DJI's FlightHub 2 management software.",[11,75,76],{},"DJI has not disclosed the total number of entries or their full geographic spread. Of the 15 named winners, the large majority are Chinese companies or university teams, with Tovar's Colombian entry among the few international names DJI highlighted — a reminder that \"global challenge\" doesn't yet mean evenly global participation.",[15,78],{},[18,80,82],{"id":81},"why-onboard-ai-matters-for-drone-operators","Why Onboard AI Matters for Drone Operators",[11,84,85,86,90],{},"Running AI on the aircraft rather than in the cloud removes the two biggest practical constraints on real-time drone analytics: network dependence and turnaround delay. A bridge inspection drone with no cell signal can still flag a crack the moment it sees one; a search-and-rescue aircraft can identify a person in open water without waiting for a link to a remote server. That shift toward on-device processing is part of a broader move across the drone industry — covered in more depth in our guide to ",[37,87,89],{"href":88},"\u002Fdrones-and-artificial-intelligence\u002F","drones and artificial intelligence"," — where autonomy and edge computing increasingly replace pilots reviewing raw footage after the flight.",[11,92,93],{},"For DJI specifically, the challenge also functions as a pipeline: winning developers get their models listed in DJI's Onboard AI Solutions Catalog and fast-tracked into its Enterprise Ecosystem partner audit, positioning their software for wider commercial distribution on DJI hardware rather than as one-off research projects.",[15,95],{},[18,97,99],{"id":98},"what-this-means-for-drone-pilots","What This Means for Drone Pilots",[11,101,102],{},"For commercial operators already flying DJI enterprise aircraft, the practical upside is a growing catalog of third-party AI models built specifically for onboard use — inspection, agriculture, and public-safety tools that install on hardware many fleets already own, rather than requiring a separate cloud subscription or custom development. Operators evaluating a Matrice 4-series or Matrice 400 purchase now have a concrete, DJI-vetted list of onboard applications to weigh against buying a drone as a bare sensor platform.",[11,104,105],{},"The competition also signals where DJI is steering its enterprise ecosystem: toward autonomous, on-aircraft decision-making rather than human review of footage after landing. Pilots and fleet managers working in agriculture, infrastructure, or emergency response should expect more DJI software updates built around this onboard-AI model over the next product cycle.",[15,107],{},[109,110],"faq",{":items":111},"[{\"question\":\"What is the DJI Enterprise Drone Onboard AI Challenge?\",\"answer\":\"It's a global DJI competition, launched in March 2026, that invites developers to build AI models running directly on DJI enterprise drones rather than in the cloud. DJI named 15 winners on August 20, 2026, across a Best Onboard AI Model category and an Industry Application Excellence category.\"},{\"question\":\"Which DJI drones support onboard AI?\",\"answer\":\"The challenge's eligible hardware includes the Matrice 4 and 4D series, Dock 3, Matrice 400, and DJI's Manifold 3 onboard computer, which supplies the processing power needed to run AI models directly on the aircraft without a cloud connection.\"},{\"question\":\"How is onboard AI different from cloud-based drone AI?\",\"answer\":\"Onboard AI processes video and sensor data on the drone itself in real time, so it works without a live network connection and returns results immediately. Cloud-based AI instead streams footage to a remote server for analysis, which adds latency and depends on connectivity that's often unreliable in the field.\"},{\"question\":\"Can developers use the winning DJI AI models commercially?\",\"answer\":\"Winning projects are listed in DJI's Onboard AI Solutions Catalog and their developers get fast-tracked into DJI's Enterprise Ecosystem partner audit, a step toward wider commercial distribution — but DJI has not announced a public release timeline for each specific model.\"}]",[15,113],{},[11,115,116],{},[117,118,119,120,126,127],"em",{},"Sources: ",[37,121,125],{"href":122,"rel":123},"https:\u002F\u002Fdeveloper.dji.com\u002Finnovation-contest\u002F",[124],"nofollow","DJI Enterprise Drone Onboard AI Challenge 2026"," | ",[37,128,131],{"href":129,"rel":130},"https:\u002F\u002Fdronexl.co\u002F2026\u002F08\u002F21\u002Fdji-onboard-ai-challenge-2026-winners\u002F",[124],"DJI Names 15 Onboard AI Challenge Winners — DroneXL",{"title":133,"searchDepth":134,"depth":134,"links":135},"",2,[136,137,138,139],{"id":20,"depth":134,"text":21},{"id":46,"depth":134,"text":47},{"id":81,"depth":134,"text":82},{"id":98,"depth":134,"text":99},"news","2026-08-23","DJI named 15 winners of its 2026 Onboard AI Challenge on Aug. 20 — drones that count crops and spot bridge cracks with AI that runs onboard, not the cloud.",false,"md","\u002Fimages\u002Fdji-matrice-4d-c6-standalone-bvlos.jpg",{},true,"\u002Fen\u002Fposts\u002Fdji-onboard-ai-challenge-2026-winners",{"title":5,"description":142},"dji-onboard-ai-challenge-2026-winners","en\u002Fposts\u002Fdji-onboard-ai-challenge-2026-winners",[153,154,155,156,157],"dji","artificial intelligence","enterprise drones","drone technology","onboard ai",null,"i4NNnIrylNe3AyhBVYa_IdW5lzjkPhf1vDJJ2kGFzT0",[161,163,173,180],{"title":5,"description":142,"date":141,"category":140,"image":145,"slug":150,"tags":162,"author":6},[153,154,155,156,157],{"title":164,"description":165,"date":141,"category":140,"image":166,"slug":167,"tags":168,"author":6},"FCC Orders 67 Drone Certification Labs to Tighten Screening","The FCC told all 67 accredited labs to tighten drone certification screening after one erroneously cleared a banned Chinese flying camera for US sale.","\u002Fimages\u002Ffcc-tcb-drone-certification-guidance-2026.jpg","fcc-tcb-drone-certification-guidance-2026",[169,170,171,172],"fcc","regulation","drone market","china",{"title":174,"description":175,"date":176,"category":140,"image":177,"slug":178,"tags":179,"author":6},"HoverAir Versa Banned in the US: FCC Pulls Approval","The FCC pulled HoverAir Versa's equipment approval days after its Indiegogo launch, halting US sales of the flying Flight Kit under Covered List drone rules.","2026-08-22","\u002Fimages\u002Fhoverair-versa-fcc-ban-2026.jpg","hoverair-versa-fcc-ban-2026",[169,170,171,172],{"title":181,"description":182,"date":176,"category":140,"image":183,"slug":184,"tags":185,"author":6},"Motorola's $1.5B Deal for Rogue-Drone Takeover Tech","Motorola Solutions closed its $1.5 billion buyout of D-Fend Solutions on Aug. 20, 2026, gaining RF cyber tech that takes control of rogue drones mid-flight.","\u002Fimages\u002Fcounter-drone-law-police-2026.jpg","motorola-dfend-counter-drone-acquisition-2026",[186,187,170,171],"counter-drone","c-uas",1787527152404]