Milestone 2
2.1: Project Plan
Computer Vision Engineer: Chiripol Sirikakan
Optics Engineer: Dorzhi Denisov
Software
An image processing algorithm will detect the location of insects by creating a movement mask and a model to predict the likelihood of a specific point being an insect. Additional models will potentially be incorporated on the video signal to detect animals and humans to disable the system.
A control loop will aim the laser at a specific point in 3D space passed by the image processing algorithm and trigger it.
Hardware
A near-infrared camera with an infrared emitter will send video to a computer, when it detects a harmful insect, it will pass a 3D coordinate t
We are currently evaluating various options for different components of the project.
Laser:
Our initial choice is a approx 0.05-watt green laser (exact wavelength to be determined) for testing the control circuit. Upon successful testing, we intend to upgrade to a 8-watt 465nm laser.
Control Circuit:
Presently, my plan involves utilizing servo motors with a closed-loop control circuit based on the rotational position. Mirrors will be attached to these motors. Additionally, we are considering the use of step motors. We will conduct tests with galvo mirrors using a closed-loop circuit with a sensor.
Project Management
The team stores all files on Google Drive and meets bi-weekly to discuss progress, impediments, and solutions.
Work Breakdown Structure
Assessment Phase (September - December 2023)
Milestone 1 (Sep 1 - Oct 11)
Identify Customer Problems and Needs
Formulate value proposition
Evaluate current market state
Identify competitors and their processes
Milestone 2 (Oct 12 - Nov 29)
Develop Project Plan
Define Roles, Responsibilities, and Assign Tasks
Conduct Research on potential Hardware and Software solutions
Formulate Product Design, Analysis, and Test Plan
Milestone 3 (Nov 30 - Dec 13)
Prepare Presentation
Assessment Phase (September 2023 - Dec 2023)
Prototype Stage (Dec 2023 - Apr 2024)
Prototype Phase (December 2023 - April 2024)
Design Planning (Dec 14 - Jan 31)
Prototype Development (Feb 1 - Feb 28)
Design Performance Evaluation
Cost Review
Verification Stage (Mar 1 - Mar 21)
Prototype Testing and Modification
2nd Prototyping Stage (Mar 22 - Apr 22)
Implement Changes
Presentation Development (Apr 23 - Apr 26)
Develop Slide Deck
Rehearse and Revise
Finalization
Innovation Expo (Apr 26, 2024)
2.2 Concepts
Acceptance Criteria
Basic Functions - The device is fully capable of accurately identify and eliminate harmful insects through the application of a precise energy laser beam
Energy - The device shall work autonomously using energy from solar panel
Image Processing - The device can differentiate between harmful and non-harmful insects, transmitting XYZ coordinates to the laser module.
Success Criteria
Our device shall effectively target and address pest problems in a non-chemical, autonomous manner. It shall be easily integrate into farms or other facilities.
Out of Scope
Removing dead insect bodies
Checking plant health
Ensuring compatibility with other devices
Assumptions
We assume our target market can afford this product.
We assume there will be no humans or animals while our device is in operation
Dependencies
We depend on third party companies to manufacture and deliver high-quality and quality-assured components.
We depend on the third party libraries for image processing.
Constraints
Our project must have an MVP ready by April 26th, 2024.
Our product can receive reimbursement of up to $400 in costs. This acts as a soft budget which can only be increased if a sponsor is found.
MVP
Raspberry Pi 4 controlled mirror for laser beam steering and an camera for detecting harmful insects.
Focus on design of fast response closed loop control system.
Focus on design image recognition software.
2.3: Concept Selection
For concept selection, we opted to use KT analysis. Initially, we assessed whether the candidate solution meets our requirements and then evaluated its final score. The justification summary for the chosen solution is provided after each table. If a cell is left blank, it means that we considered the attribute to be irrelevant to the solution under discussion.
Justification:
Despite the lower score of the Raspberry Pi, we chose it over Arduino and microcontrollers due to the necessity of image processing on local machine. The Raspberry Pi's superior computational power is indispensable for this task, making it the only viable solution that satisfies all three of our needs.
Justification:
We chose laser technology over water and salt/sand-based methods due to its superior accuracy, which is unaffected by wind or other environmental factors.
Justification:
While galvo (galvanometer) systems are more expensive than servo or stepper motor solutions, their superior accuracy makes them the preferred choice for our application.
Justification:
While using a sonar system would be the most cost effective solution, the accuracy would suffer greatly, additionally a normal camera costs about the same but the accuracy is not as good.
2.4: Design
2.5: Analysis
Hardware Specifications
The hardware in our project involve these components:
Raspberry Pi 4: A microcontroller used to integrate the control loop, camera, laser and run image processing algorithms.
Infrared Camera: Used to detect insects. Requires less image processing in comparison to other cameras.
Galvo Mirrors: Used to steer the laser beam.
Laser: Used for eliminating harmful insects.
Software Specifications
The software is mainly reliant on these following technologies
Python: programming language that will integrate with OpenCV to be able to detect different types of insects.
OpenCV: open source image processing library that gives the team various ways to perform image processing.
2.6: Test Plan
Laser Testing:
The optics engineer will initially test a 0.1W laser system for safety reasons, subsequently transitioning to a 2W laser. The testing procedure involves placing the laser system at a known distance from a marked wall with known coordinates. Inputting this data into the laser system, it should accurately point to the marked positions on the wall. Evaluation criteria include accuracy, steady-state error, and response time.
Computer Vision Testing:
The computer vision engineer will conduct two types of tests. The first assesses the accuracy of coordinates, with the computer vision system placed at a known distance from a marked wall. The system should output the corresponding coordinates. The second test evaluates the system's ability to identify insects. In a controlled environment with insects, the computer vision system should draw boxes around each insect, labeling them accordingly.
System Testing:
The laser system and computer vision system will be integrated for comprehensive system testing. A testing environment (container) containing insects will be used, and the system is expected to successfully eliminate all insects within the test environment.