Jeffrey T' Drake

1
Bulk Insect Sorting Using Image Analysis

• Jeffrey T. Drake

Acknowledgments USDA-APHIS-PPQ Center for Plant
Health Science and Technology Dan Fieselmann
USDA Forest Service Forest Health
Protection/Technology Enterprise Team Frank
Sapio New Mexico State University Biological
Control Laboratory Dr. Joe Ellington Tracey
Carrillo
2
Objectives Automated Insect Survey

• Surveys Overwhelmed by Volume
• Backlog
• Sub-Optimal Sampling
• Support Insect Survey
• Develop Tools to Aid in Rapid Screening
• Remote Sensing (distance in inches)
• Physical Sorting
• Incremental Approach
• Foundations (Proof-of-Concept)
• Extensions

3
Background Automated Insect Survey
Support Insect Survey Via Automation

• Proof-of-Concept
• Cotton
• Complex-Diverse
• Understood

4
Background Automated Insect Survey

• Human Genome Project

Genes Mapped
1943 Discovery
1993 Automation
2000
2003
5
Background Automated Insect Survey
Survey Requirements - CAPS
6
Technology Automated Insect Survey
Decision
Input
Feature Extraction
Sensing
Segmentation
Classification
7
Capabilities Automated Insect Survey
8
Workflow Automated Insect Survey
9
Extensions Automated Insect Survey
Applications Bark Boring Beetles Forest Health
Surveys Fruit Fly Control

• Species Level ID

10
Extensions Automated Insect Survey
Employing Techniques for Limited Species ID
11
Extensions Automated Insect Survey
Evaluating Techniques for Limited Species ID
Laser Topology Mapping
12
Extensions Bark Boring Beetles
Separate the wheat from the chafe

• Requires Species Level ID

13
Extensions Industrial Robotics
Requires Physical Sorting
Study Doable with Existing Commercial Systems
14
Extensions Bulk Sorting Workstation
15
Extensions Bulk Sorting Workstation
Insect Sorting Robot
16
What This Can Mean For You
Survey
Biocontrol
Insect Densities Fast Reliable