Vision-Guided Plant Trimming to Protect Flowers and Trichomes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mechanical and computer-vision based trimming technologies for cannabis and other plants are inefficient, prone to over-trimming, and damage to valuable flower and trichomes, due to variations in plant shapes, colors, and structures, and lack precise control.
Innovation Solution
An automated robotic system with a transport mechanism, cutting support mechanism, camera, imaging stage, and central processor that uses image processing and machine learning to create a trim map, guiding a cutting armature to precisely identify and remove low-cannabinoid leaves while protecting the high-cannabinoid flowers and trichomes, employing sensors for plant water content and physical resistance data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If mechanical trimming devices are used to automate the trimming process, then productivity is improved, but manufacturing precision deteriorates causing over-trimming and damage to flower and trichomes
Solution Approach 1:
The patent replaces traditional mechanical trimming devices with a robotic system that uses computer vision and machine learning to guide precise mechanical cutting. The system captures images of the plant, processes them through AI algorithms to identify leaves versus flowers, and then directs a robotic cutter to trim only the identified leaves, substituting purely mechanical automation with an intelligent hybrid system that maintains both speed and precision.
Solution Approach 2:
The system creates a digital copy or map of the plant structure through image capture and processing. This digital representation allows the system to analyze plant anatomy, identify trim targets, and plan cutting paths before execution, enabling precise trimming without direct mechanical contact during the identification phase.
2Ease of operation
If traditional mechanical trimmers are used, then ease of operation is improved through automation, but object-affected harmful factors increase due to damage to trichomes and flowers
Solution Approach 1:
The system replaces blind mechanical trimming with vision-guided robotic trimming. Computer vision cameras capture detailed images of the plant, machine learning algorithms identify the anatomical structures, and the robotic system executes precise cuts only where needed, eliminating the random damage caused by traditional mechanical trimmers while maintaining automation benefits.
Solution Approach 2:
The system implements a feedback loop where images of the plant are captured, processed to identify leaves and flowers, and used to guide the trimming action in real-time. This closed-loop control ensures that cutting actions are based on actual plant anatomy observations, preventing damage to valuable structures while maintaining automated operation.
3Manufacturing precision
If computer-vision systems are implemented to improve trimming precision, then manufacturing precision is improved, but device complexity increases
Solution Approach 1:
The system employs a universal image processing pipeline that handles various plant types and configurations through a single machine learning model. The same camera and processing system can identify different leaf shapes, sizes, and positions across diverse plant specimens, reducing the need for multiple specialized devices while maintaining high precision across applications.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The automated trimming of an untrimmed plant stem comprises a device for transporting an untrimmed plant stem to an imaging stage, using a camera to capture at least one image of the plant stem, sending the at least one image to a central processor which creates a trim map which constructs a pattern for moving a cutting armature along the plant stem, and trimming the plant stem by cutting structure of the cutting armature to change the untrimmed plant stem to a trimmed plant stem.