Appraisal of Low-Cost Pushbroom Hyper-Spectral Sensor Systems for Material Classification in Reflectance.

Area of Science:

• Geoscience
• Planetary Science
• Remote Sensing

Background:

• Near-infrared (NIR) remote sensing is crucial for analyzing vegetation and geological features.
• Extra-terrestrial applications include mineral identification on the Moon, Mars, and asteroids within the 600-800 nm range.
• Advancements in sensor and processor technology enable the development of affordable scientific instruments.

Purpose of the Study:

• To evaluate the utility of Raspberry Pi cameras and a panchromatic astronomy camera for measuring and processing spectral reflectance.
• To assess the feasibility of using low-cost cameras for mineral identification in remote sensing applications.
• To determine if common STEM cameras can achieve scientific results in spectral analysis.

Main Methods:

• Two types of Raspberry Pi cameras and one panchromatic astronomy camera were integrated into a pushbroom sensor.
• Spectral reflectance measurements were taken for 15 test materials with phenomenology between 600 and 800 nm.
• Calibration against a spectrometer accounted for sensor effects, image processing, and compression.

Main Results:

• Algorithmic classification successfully identified all 15 test materials based on their spectral properties.
• Color Raspberry Pi cameras, contrary to expectations, effectively recorded and distinguished between most minerals.
• These cameras utilized secondary infrared transmissions in the Bayer filter, broadening their effective spectral range.

Conclusions:

• Low-cost Raspberry Pi cameras are suitable for mineral identification in NIR remote sensing, particularly for space exploration.
• The use of Bayer filters in standard cameras does not preclude their utility for spectral analysis.
• While removing the Bayer filter could enhance sensitivity, it may not be essential for many applications.