Agricultural Row Guidance Using 3D Camera Image Masking

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Agricultural working machines face challenges in accurately determining the position of plant rows and soil sections due to camera-generated images that include machine parts, leading to incorrect image recognition and reduced field processing quality.

Innovation Solution

Implementing an image processing unit that dynamically or statically hides image areas containing machine parts, allowing only soil and plants to be considered in image processing, using a 3D camera and image mask to adjust for changes in camera position and orientation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If camera captures are used to generate position information, then automation and monitoring capabilities are improved, but image recognition accuracy deteriorates due to capture of machine parts

Engineering Contradiction:
Improveautomation of monitoring and controlVSAvoidposition information accuracy
Core Design Contradiction:
Extent of automationVSMeasurement precision

Solution Approach 1:

The patent extracts and removes machine parts from the image data through masking before position information is generated. The image processing unit creates masks that exclude regions containing tractor and agricultural work unit parts, ensuring only relevant field information is processed for accurate position detection.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces an intermediary masking mechanism between the camera and image processing stages. This mediator (the mask) selectively filters out harmful data (machine parts) while preserving useful data (field information), enabling accurate position information generation despite the presence of automated monitoring systems.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Quantity of substance

If image processing includes all captured areas, then completeness of data is improved, but processing accuracy deteriorates due to inclusion of machine parts

Engineering Contradiction:
Improveamount of image dataVSAvoidposition information precision
Core Design Contradiction:
Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent applies different quality treatment to different regions of the image data. Through localized masking, specific regions containing machine parts are excluded from processing while other regions (field areas) are processed normally. This selective approach maintains processing accuracy without discarding valuable field information.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3818799A1Agricultural machine
Publication Date: 2021.05.12 CLAAS E SYSTEMS GMBH
  • EP3818799A1 patent drawingFigure 1
  • EP3818799A1 patent drawingFigure 2a~2b
  • EP3818799A1 patent drawing

AI summary

The invention relates to an agricultural machine comprising a tractor (1), in particular a tractor (3), with at least one agricultural working unit (2) for cultivating a field crop comprising a plurality of crops, and with a control device (5), wherein the control device (5) comprises a camera arrangement (6) with at least one camera (8), in particular a 3D camera, which generates images of the field crop by means of the camera (8), wherein the control device (5) comprises an image processing unit (7) which, by means of image processing based on the generated images, generates position information about the position of plant rows (9) and/or soil sections (10) to be cultivated, wherein the control device (5) adjusts the lateral position of the agricultural working unit (2) based on the generated position information.It is proposed that the image processing unit (7) be configured to dynamically and/or statically hide image areas (A) in the generated images that are to be disregarded during image processing.