Optical Sprayer Control for Real-Time Targeted Field Application
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current agricultural sprayers and applicators apply materials uniformly across fields without considering localized needs, leading to inefficiencies and environmental impact, and existing prescription-based methods are inaccurate due to outdated data and image quality issues.
Innovation Solution
An agricultural applicator equipped with on-board real-time image sensors and controllers that identify targets for precise application of materials, adjusting nozzle operation based on real-time imaging and environmental conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If broadcast application is used to apply material uniformly across the field, then coverage is ensured, but material waste increases and environmental impact worsens
Solution Approach 1:
The system transitions from uniform broadcast application to localized targeted application by using image sensors to identify specific weed locations and controlling individual nozzles to apply material only at those locations, thereby reducing overall material usage while maintaining effective coverage where needed
Solution Approach 2:
The spray boom is divided into multiple independently controllable nozzle sections, allowing selective activation of individual nozzles based on real-time weed detection, enabling precise material application only where targets are present
2Loss of substance
If prescription-based application is used with pre-generated maps, then material application is reduced, but accuracy deteriorates due to outdated data
Solution Approach 1:
The system performs preliminary weed identification and mapping during the current pass through the field using real-time image sensing, creating an updated prescription map that is immediately actionable, eliminating the time delay and data staleness associated with pre-season scouting and drone surveys
Solution Approach 2:
The system continuously captures real-time images of the field, processes them to identify weeds, and immediately adjusts nozzle operation based on current conditions, creating a closed-loop feedback system that ensures high accuracy in target location and material application
3Measurement precision
If real-time image sensing is implemented for precise target identification, then material application accuracy improves, but device complexity increases
Solution Approach 1:
The image sensor system serves multiple functions: it identifies weed targets, maps their locations, guides nozzle operation, and creates prescription data, allowing a single sensing system to perform what would otherwise require multiple separate subsystems
Solution Approach 2:
The controller acts as an intermediary that processes image data from the sensor and translates it into nozzle control signals, simplifying the interface between the sensing system and the application mechanism while enabling precise coordination
4Ease of manufacture
If traditional scouting methods are used to generate prescriptions, then data collection is simple, but data freshness and accuracy deteriorate
Solution Approach 1:
The system performs its own data collection and analysis autonomously using onboard image sensors and processing capabilities, eliminating the need for separate manual scouting operations or external drone surveys while maintaining continuous data freshness throughout the application process
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
An agricultural machine (100) that travels across a field in a direction of travel is disclosed. The agricultural machine (100), comprising: a material reservoir; a controllable valve (190); a pump (184) that pumps material from the material reservoir to the controllable valve (190); a plurality of optical sensors (122) that each capture an image, of a plurality of captured images, of a portion of the field ahead of the controllable valve (190) in the direction of travel; a processing module (124) that receives the plurality of captured images and includes a target identification system (158) that identifies a target in at least one image of the plurality of captured images; and a valve controller (170) that generates a valve control signal to control the controllable valve (190) to apply the material to the target. Furthermore, a method of controlling such agricultural machine (100) is disclosed.