Automated Plant Sorting via Computer Vision and Machine Learning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The strawberry industry relies heavily on manual labor for sorting plants, which is time-consuming and costly, and existing automated systems are not effective for bare-root crop sorting, leading to variable quality and inefficiencies in plant distribution.

Innovation Solution

An automated system utilizing advanced computer vision and machine learning algorithms to classify and sort strawberry plants into quality grades, incorporating pixel-based evaluation and real-time image processing to direct plants into specific bins for sale or rejection, with adaptable features for changing crop conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual labor is used to sort plants, then quality assessment can be performed, but the process is time-consuming and costly

Engineering Contradiction:
Improvequality assessment accuracyVSAvoidsorting speed
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces the manual mechanical sorting system with an automated computer vision system. Cameras capture images of plants, and machine learning algorithms automatically classify and sort plants based on visual characteristics, eliminating the need for human labor while maintaining quality assessment accuracy and significantly increasing sorting speed.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system creates digital copies (images) of the plants and analyzes these copies using machine learning algorithms. Instead of physically handling each plant, the system captures visual information and processes it computationally to make sorting decisions, which is both faster and more consistent.

Inventive Principle:
Principle #26Copying

2Productivity

If automated systems are used for plant sorting, then sorting speed increases, but existing systems are not effective for bare-root crop sorting

Engineering Contradiction:
Improvesorting speedVSAvoidsorting effectiveness
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent adapts the computer vision system specifically for bare-root crops by adjusting the machine learning algorithms and image processing parameters to handle the unique characteristics of bare-root plants. This includes modifying the classification models to recognize specific root structures and plant features relevant to bare-root crops, ensuring reliable sorting while maintaining high speed.

Inventive Principle:
Principle #35Parameter changes

3Reliability

If manual sorting is used, then quality control is maintained, but the time plants spend out of the ground increases

Engineering Contradiction:
Improvequality controlVSAvoidtime out of ground
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The automated computer vision system performs quality control faster than manual sorting. By using high-speed cameras and machine learning algorithms, the system can assess and sort plants in seconds, significantly reducing the time plants spend out of the ground while maintaining consistent quality control through standardized digital assessment criteria.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

4Productivity

If hundreds of migrant workers are employed for sorting, then large volume of plants can be sorted, but labor costs increase

Engineering Contradiction:
Improvevolume of plants sortedVSAvoidlabor requirements
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent replaces the complex human labor system with an automated technological system. Instead of employing hundreds of migrant workers, the system uses cameras, image processing software, and machine learning algorithms to perform sorting, dramatically reducing labor requirements while maintaining the capability to sort large volumes of plants efficiently.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP2548147B1Method to recognize and classify a bare-root plant
Publication Date: 2020.09.16 CARNEGIE MELLON UNIV
  • EP2548147B1 patent drawingFigure 4
  • EP2548147B1 patent drawingFigure 5
  • EP2548147B1 patent drawingFigure 9A~9B

AI summary

The present invention encompasses software that brings together computer vision and machine learning algorithms that can evaluate and sort plants into desired categories. While one embodiment of the present invention is directed toward strawberry plants, the software engine described is not specifically designed for strawberry plants but can be used for many different types of plants that require sophisticated quality sorting. The present invention is a sequence of software operations that can be applied to various crops (or other objects besides plants) in a re-usable fashion.