Target definition
The system must define what qualifies for treatment, what must never be treated, and how unknown, hidden, overlapping, or ambiguous objects are handled.
DETECT THE TARGET, CONTROL THE NOZZLE, VERIFY THE RESULT
Precision Spot
Spraying
Precision spot spraying closes a fast physical loop from target observation to nozzle action. Its usefulness depends on detection, classification, location, latency, machine motion, plumbing response, spray quality, label compliance, and outcome verification.
5CHAPTERS5VISUAL BLOCKS4GRAPH LINKS3SOURCES
ACT / SELECTED CONCEPTPrecision Spot SprayingStart with the role, then move through the editorial sequence.
HOW TO READ THIS PAGE
Visual explanationA diagram or operating scene makes the relationship visible.
Structured modelA flow, comparison, capability set, or boundary map organizes the idea.
Guided explanationOriginal prose connects the concept to its operating context.
01 / Overview
Recognition is only the first
half of targeted application.
A targeted sprayer may use real-time sensing, a prior map, or a combination. The decision layer must convert a qualified target into a legal and agronomically appropriate action while accounting for machine speed, sensor-to-nozzle offset, latency, boom motion, fluid dynamics, and coverage.
USDA ARS describes a precision specialty-crop system that uses LiDAR-derived canopy information to control variable nozzle flow, and a separate research project that evaluates optical sensing and UAV weed mapping for variable herbicide application. Those cases illustrate different target definitions rather than one universal architecture.
02 / Control loop
See, decide, align,
spray, prove.
OBSERVECrop and field sceneCrop, target, background, growth stage, lighting, weather, speed, occlusion, field and operating context
DETECTSensor and modelSensor identity, view, calibration, software and model version, confidence, unsupported conditions, latency, and outputAUTHORIZEApplication decisionQualified target definition, product label and local authority, thresholds, exclusions, buffers, stop conditions, and supervisionALIGNMachine timing and geometryPosition, heading, sensor-to-nozzle offset, boom movement, speed, delay, section or nozzle identity, and predictionACTNozzle and fluid responseCommand, valve, pressure, flow, nozzle, plumbing, product properties, dynamics, reported state, and physical uncertaintyREVIEWApplication evidenceAs-applied record, missed and treated context, alerts, interruptions, physical checks, efficacy and safety evidence, and unresolved questions- 01Crop and field scene→Sensor and modelbounded observations
- 02Sensor and model→Application decisioncandidate target evidence
- 03Application decision→Machine timing and geometryqualified bounded action intent
- 04Machine timing and geometry→Nozzle and fluid responsetime- and location-specific command
- 05Nozzle and fluid response→Application evidencereported actuation and field question
- 06Application evidence←Sensor and modelvalidation evidence and failure cases
01PERCEIVE / 01Observe the target zoneCrop, weed, canopy, soil, lighting, dust, residue, speed, sensor health, and position
02QUALIFY / 02Make a bounded treatment decisionClass, confidence, target rule, product label, threshold, exclusion, and fallback
03ALIGN / 03Transform sensing into nozzle timeCalibration, geometry, latency, travel speed, boom motion, valve delay, pressure, and section state
04VERIFY / 04Record and inspect the outcomeCommand, actual flow, coverage, miss, false application, drift condition, refill, cleaning, and field result
03 / System layers
An accurate model can still
produce a misplaced spray.
Spatial and temporal alignment
Sensor calibration, vehicle motion, boom movement, terrain, computing delay, communication, valve response, and fluid travel determine placement.
Application physics
Nozzle, pressure, flow, droplet spectrum, product properties, weather, canopy, travel speed, mixing, and wear determine delivery beyond the digital command.
Independent verification
Confusion errors, missed targets, unintended treatment, actual flow, coverage, crop injury, control outcome, repeatability, and operating domain require field evidence.
04 / Operating modes
Targeted application includes
several different control problems.
ModeDecision basisPrimary boundary
Map-based spot treatmentPrior georeferenced target or treatment mapTarget movement, map age, georegistration, and field change between survey and spray
Real-time green-on-brown detectionVegetation distinguished from bare backgroundCannot by itself separate crop from weed where both are green
Real-time crop-weed classificationMachine-vision model assigns target class or maskCrop stage, species, overlap, residue, light, novelty, and confidence affect errors
Canopy-adaptive sprayingCanopy presence, size, shape, or density controls deliveryCanopy measurement and dose logic are crop-, system-, and validation-specific
05 / Boundaries
Input reduction does not prove
control, safety, or profit.
The pesticide label remains authoritative.Target, crop, rate, carrier, nozzle, pressure, timing, buffer, weather, personal protection, cleaning, records, and local law remain product- and jurisdiction-specific.
False negatives and false positives have different costs.A missed target, crop hit, off-target application, delayed control, resistance risk, and unnecessary treatment require separate evaluation rather than one accuracy score.
Research results are configuration-specific.Reported outcomes belong to the tested crop, target, sensor, model, sprayer, nozzle, product, weather, speed, field, and evaluation protocol.
06 / Connections
See the system around this concept.
Follow incoming and outgoing relationship records to understand what supplies, informs, enables, coordinates with, or extends this technology in the published knowledge graph.
Relationship radar / published edges4 records / 4 neighboring systems
Incoming03records point toward this concept
act rolePrecision Spot SprayingSelected technology
Outgoing01records point from this concept
04connections visible
observe / Machine perceptionAgricultural Machine Vision ↗can supply target perception to
A validated machine-vision system can supply target class, location, mask, or confidence information to a real-time spot-spraying controller.
observe / Aerial platformsAgricultural Unmanned Aircraft Systems ↗can provide prior target maps to
A qualified aerial weed-mapping workflow can contribute prior spatial target evidence to a later map-based spot application.
act / Field applicationVariable Rate Technology ↗is a target-bounded application mode that can coordinate with
Spot treatment and broader rate variation solve related but distinct spatial application decisions and may share positioning, control, and as-applied records.
act / Application controlPulse-Width Modulation Spray Control ↗can provide nozzle-level actuation for
A qualified PWM system can actuate supported nozzle-level target decisions, but perception, classification, treatment authority, spatial alignment, valve and fluid response, misses, unintended application and outcome verification remain separate responsibilities.
3CONNECTED ROUTES2–4STEP POSITIONS32ROUTE SOURCE LINKS
Operating practice
From crop signal to targeted spray
Separate broad crop sensing, mapped aerial evidence, real-time perception, nozzle-level action, and practical sprayer verification in one bounded learning route.
- 01 / SIGNALRead the crop canopyTechnology→
- 02 / MAPCollect spatial evidence from aboveTechnology→
- 03 / PERCEIVEClassify the target in real timeTechnology→
- 04 / ACTAlign the target with the nozzleTechnology→
- 05 / MODULATETranslate a target into nozzle timeTechnology→
- 06 / PROVEVerify the sprayer before treatmentField guide
CURRENT POSITION04
04 / ACT
Open the complete route ↗Align the target with the nozzle
Connect target logic to sensor offset, latency, machine motion, valve response, spray physics, label constraints, and records.
07 / Evidence
Primary sources.
This briefing uses USDA ARS material on LiDAR-guided specialty-crop spraying, optical sensing and weed mapping, and machine vision. It does not reproduce product algorithms, recommend pesticides, or transfer research performance to untested systems.
01
Precision Sprayer Benefits Growers and the EnvironmentUSDA Agricultural Research Service · Accessed 2026-07-20
02
Predicting Nitrogen and Sulfur Deficiency in Corn using Optical Sensors and Weed Mapping with UAV Multispectral SensorsUSDA Agricultural Research Service · Accessed 2026-07-20
03
Robust crop and weed segmentation under uncontrolled outdoor illuminationUSDA Agricultural Research Service · Accessed 2026-07-15