Plastic Scanner — building an open-source NIR sensor

While working on plastic recycling in early 2023, I wanted to explore practical ways to identify and sort different plastics. I built and tested a sensor board based on the open-source Plastic Scanner Project, which investigates low-cost near-infrared sensing for plastic identification.

2023

My role: Electronics assembly, firmware work, Python tools, testing and hardware debugging

Based on: The open-source Plastic Scanner Project

A working measurement setup

My build focused on assembling, testing and debugging the measurement hardware and its software interface, creating a working setup for exploring the sensor readings from different samples. The wider Plastic Scanner Project aims to use sensor data and machine learning to identify plastic types.

The setup used eight infrared LEDs and a broadband detector to collect reflected-light measurements at wavelengths spanning approximately 850–1,650 nm. Plotting software made it possible to inspect the measurements from different samples.

Scanner connected to a laptop with labelled plastic samples on the workbench — Scanning different plastic samples

Plastic Scanner plotting interface displaying example data in dummy-data mode — Plotting interface with example data

From components to a working sensor

I sourced the boards and components, then assembled the surface-mount circuitry using a hot-air soldering station. Building the board in functional groups allowed me to test each section before moving on. The completed sensor board connected to an Arduino Uno.

Testing a partially assembled sensor board against its schematic — Testing each circuit section

Completed sensor circuit board and component list — Completed board assembly

Open scanner enclosure showing the sensor board and Arduino — Inside the scanner

Assembled scanner in its closed enclosure — Assembled scanner

Tracking down inconsistent readings

The first board produced intermittent results. After checking the software, I turned to the electronics. I found issues in the schematics and LED current-limiting resistor values, then traced the intermittent connection to a cold solder joint on an IC pin.

Sensor circuit board under microscope illumination on the workbench — Sensor board under the microscope

Microscope and screen showing an enlarged view of the circuit board — Inspecting the PCB under a microscope

Probing the sensor board connected to an Arduino Uno during hardware debugging — Debugging the sensor board

A second board, assembled with particular attention to those connections, resolved the intermittent behavior. The project gave me practical experience in surface-mount assembly, optical sensing and systematic hardware debugging.

↑Back to Top