Spectrum Acquisition System Using Photoelectric Detection and Its Data Mining Simulation of Spectral Texture Features

Abstract

Weak spectral signal detection technology is a new subject in the measurement engineering field. It obtains the characteristic information of the measured object based on the characteristics of light diffraction, interference, polarization and so on. Regarding light source entropy, the artificial light source of mercury argon lamp with a spectrum of 250 nm~918 nm is selected as the first research object. The grating monochromator with built-in grating scale lines (1200 g/mm, 600 g/mm) is selected as the light splitting device. Due to the weak spectral signal, the output signal is vulnerable to noise interference after a certain amount of photoelectric conversion. AD202 variable voltage coupling filtering technology is adopted. It means that in the signal isolation and amplification process, the AD202 variable voltage coupled filter amplifier chip is employed to design the current/voltage conversion circuit and the same direction proportional operation amplifier circuit. Meanwhile, a second-order low-pass filter circuit is designed based on the characteristics of the OP07 operational amplifier. For the data mining of multispectral image texture features, Contourlet transform is used, that is, the multispectral image texture features are fuzzy preprocessed to remove the adjacent range correlation knowledge. On this basis, the texture features of multispectral images are analyzed. In the experiment, Contourlet transform is adopted to transform multispectral images from spatial domain to frequency domain, which has the advantages of short extraction time, low cost, and high mining accuracy. A spectrum acquisition system based on photoelectric detection is designed. With chlorophyll as the research object, the sample solution to be tested has high photoelectric absorbance in the range of blue-violet light (400 nm~500 nm) and red-orange light (600 nm~700 nm).