Recompose: Cognitive Composting

About the Project

Every day, millions of pounds of food are thrown away, contributing to massive environmental and economic costs. 30-40% of all food intake in America is thrown out or wasted, while food waste across the globe contributes to 8-11% of greenhouse gas emissions.

At Princeton Day School in New Jersey, an average of 157 pounds of food is wasted every week. Across the U.S., the average student discards 67 pounds yearly, translating to 18,000 to 33,000 pounds of waste per school each year.

30-40%

Food Waste in America

8-11%

Global GHG Emissions

157 lbs

Weekly Waste at PDS

Research Questions

Computer Vision Accuracy

How accurately can CV-based depth-sensor models predict food waste weight compared to traditional density-based assumptions?

Real-Time Impact

How does displaying real-time Carbon Dioxide, Methane, water, and energy loss metrics influence students' food waste behavior?

Behavioral Psychology

How do various psychological strategies influence food waste reduction, and does layering multiple strategies produce additive effects?

Methodology

The project utilizes a depth camera to capture RGB-D photos, which are processed to identify food waste. The workflow is as follows:

Depth Camera: Takes RGB-D photos of food items.
Motion Detection: Uses Intel RealSense Camera to detect motion.
Image Processing: OpenCV creates a bounding box around detected items and crops the image.
Data Parsing: Isolates depth and RGB data for further analysis.
Volume Calculation: Uses linear algebra to calculate object volume.
Output: Data is displayed on-screen and saved in a CSV file for further analysis.

This methodology allows for real-time feedback and accurate measurement of food waste, contributing to sustainable food management.

Environmental Impact

Carbon Reduction

Real-time tracking of CO₂ and methane emissions

Water Conservation

Monitoring water usage equivalent to food waste

Energy Efficiency

Tracking energy loss from food waste

Introduction

Food waste makes up 30-40% of the U.S. food supply and contributes to 10% of global emissions.

Food waste is expected to reach 2.1 billion tons by 2030, leading to rising temperatures, extreme weather, and biodiversity loss.

Engineering Goals

Detect Food Waste: The system automatically identifies and measures food waste.
Calculate Impact: The system estimates weight and shows how much CO₂, water, and energy is wasted.
Encourage Change: A screen gives real-time feedback to help people waste less food.

Statistics and Results

Food waste reduction: 36.9% per day (55 lbs decrease).

Estimated impact at Princeton Day School: 9,900 lbs of food wasted annually.

Discussion

This system demonstrated measurable, statistically significant impact on food waste reduction.

Future plans include implementing the device in more schools and expanding its capabilities.

Psychology and Behavioral Insights

This section explores the psychological strategies used to influence food waste behavior.

Real-Time Feedback: Displaying real-time metrics to influence behavior.
Behavioral Nudges: Using subtle cues to encourage less waste.
Social Influence: Leveraging peer pressure to reduce waste.
Gamification: Introducing game-like elements to make waste reduction engaging.

These strategies are designed to create a positive impact on food waste reduction by engaging students in a meaningful way.

Example Interface for Project

This interface provides a detailed breakdown of the environmental impact of waste.

Waste Amount

You wasted 0 pounds of food.

Carbon and Methane Emissions

You produced 0 kg of CO₂ and 0 kg of methane emissions.

Environmental Impact

Your waste destroyed 0 trees and contributed to the destruction of 0 habitats.

Water and Energy Waste

Wasted 0 gallons of water and cost 0 hours of energy.

Help Us Reach Our Goal

Help us waste less than 50 pounds as a school by wasting 0 pounds less.