Spacing-Aware Plant Detection for Crop-Weed Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine vision systems in precision agriculture are prone to errors in distinguishing between crops and weeds, leading to reduced task quality or crop damage due to misclassification, especially when weeds and crops have similar appearances and growth phases.
Innovation Solution
A spacing-aware plant detection model that utilizes the average inter-crop spacing to bias the plant detection process, enhancing the accuracy of identifying weeds by expecting crops to be in specific zones and adjusting classification probabilities based on this spacing, thereby improving task execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine vision systems are used to detect and classify plants, then automated agricultural tasks can be performed, but misclassification errors occur when weeds and crops have similar appearances
Solution Approach 1:
The system changes the parameter used for plant classification from purely visual appearance to include spatial positioning information. By incorporating the detected position of plants relative to expected crop locations, the system adjusts classification probabilities to reduce misclassification errors between visually similar weeds and crops
Solution Approach 2:
The system introduces an intermediary layer between image capture and final classification. A plant detection model first identifies potential plants and their positions, then this positional information serves as an intermediary that biases the final classification decision, allowing the system to resolve ambiguities when visual features alone are insufficient
2Speed
If machine vision systems classify plants based on appearance alone, then classification can be performed quickly, but errors increase when weeds and crops look alike at different growth phases
Solution Approach 1:
The system augmentsthe classification process by adding spatial position as an additional parameter. Instead of relying solely on appearance-based classification, the system incorporates the location of detected plants within the crop row to inform classification decisions, thereby improving accuracy without significantly impacting processing speed
Data Source
AI summary
Methods and systems for controlling robotic actions for agricultural tasks are disclosed which use a spacing-aware plant detection model. A disclosed method, in which all steps are computer-implemented, includes receiving, using an imager moving along a crop row, at least one image of at least a portion of the crop row. The method also includes using the at least one image, a plant detection model, and an average inter-crop spacing for the crop row to generate an output from the plant detection model. The plant detection model is spacing aware in that the output of the plant detection model is altered or overridden based on the average inter-crop spacing. The method also includes outputting a control signal for the robotic action based on the output from the biased plant detection model. The method also includes conducting the robotic action for the agricultural task in response to the control signal.