Agricultural Spray Nozzle Control Using Overlapping Plant Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing agricultural spray technologies face challenges in accurately classifying crop plants and weeds due to incorrect classification at the edge of the view field, leading to inefficient application of spray agents.
Innovation Solution
The method involves detecting overlapping field sections using optical detection units with inclined optical axes, classifying plants in image evaluation areas with defined overlaps, and applying spray agents based on plant identification numbers determined within these areas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single optical detection unit is used to capture the field section, then the device complexity is reduced, but measurement precision deteriorates due to incorrect classification of plants at the edge of the view field
Solution Approach 1:
The field of view is divided into multiple evaluation areas (first image evaluation area and second image evaluation area) corresponding to different optical detection units. Each unit captures and classifies plants in its specific evaluation area, with the overlapping region allowing verification and correction of edge cases. This segmentation resolves the contradiction by distributing the classification task across multiple units while maintaining overall system manageability.
2Measurement precision
If multiple optical detection units with overlapping fields of view are used, then measurement precision is improved by reducing edge classification errors, but device complexity increases
Solution Approach 1:
The classification results from multiple optical detection units are merged in the overlapping evaluation areas. The control unit compares classifications from different units for plants in the overlapping region and resolves discrepancies, thereby improving overall classification accuracy while using a manageable number of detection units.
3Productivity
If the depth of image evaluation areas is increased to cover more plants, then productivity is improved by treating more plants per detection, but manufacturing precision deteriorates due to increased classification errors at area boundaries
Solution Approach 1:
The total evaluation depth is segmented into multiple smaller evaluation areas with controlled depths. Each area is classified by a dedicated optical detection unit, reducing boundary classification errors. The overlapping regions between areas allow for verification and correction, maintaining precision while achieving comprehensive coverage through the combined depth of multiple areas.
Data Source
AI summary
A method for applying a spray agent to an agricultural land using a spray nozzle unit of an agricultural spray device. The method includes detecting first and second field sections of the agricultural land using an optical detection unit to obtain first and second pieces of image information from the first and second field sections, wherein the pieces of image information have a common lateral overlapping area having a depth in the travel direction of the agricultural spray device; classifying plants in the obtained pieces of image information using a control unit; determining first and second plant identification numbers for first and second image evaluation areas in the obtained first and second pieces of image information using the classified plants; and applying the spray agent as a function of the determined plant identification numbers using the spray nozzle unit of the agricultural spray device.