LED Spectrum Measurement Using Low-Cost Spectrum Sensor Array and Particle Swarm Optimization

A LED spectrum measurement system using a low-cost spectrum sensor array is proposed in this work. We note that since the spectral transmittance properties of the spectrum sensors are off ideal, a robust mapping process to transform the sensor outputs to the information of LED spectrum is demanded. For mapping the raw data from the low-cost spectrum sensor array to the spectrum component of LED lights, particle swarm optimization (PSO) with full factorial design (FFD), namely FFD PSO, is utilized to search the optimal center locations and the widths for the Gaussian basis functions in use. As a result, we show that the achievable measurement accuracy is competitive to a professional optical spectrometer.