A Hyperspectral Image Detection System Based On Acousto-optic Tunable Filter

In order to confirm the the freshness of meat, a hyperspectral image detection system based on acousto-optic tunable filter had been studied and designed.The high/multispectral imaging s of samples can be achieved under the 550 nm to 1000 nm wavelength.The system build up a goog basis for the research of meat freshness based on hyperspectral imaging system.The selection of hardwares based on machine vision theory, including the light source, lens, camera equipment, etc. Software of the system is based on VC + + and HALCON Softwares, with a simple user interface which is easy to understand and operate. Software programming realized functions of monitoring real-time images, controling parameters, storaging data and preprocessing.Operaters can change the condition of wavelengths according to specific requirements by changing the configuration file and reset parameters of bitting, exposure time, deep of images and gathering model on the interface.Calibration and other evaluation had been carried out on the system.The tests of three days of continuous error test, light intensity test, the system stability test and signal-to-noise ratio adjustment test show, the frame rate of camera is 16.2 frames per second and the frame rate of system is 8.2 frames per second under the condition of 1344 * 1024 pixels, which means the system owns a quick acquisition speed and a high frame rate. the rescher of thermal effect mechanism of Lighting system to samples show when even three halogen lamps working at the same time the relative influence is 0.673 compared with the environmental impact in 25℃. Repetitive error of the brightness, according to the spectral image acquisition of three continuous days, is ±1% in the ROIs.Based on designing hyperspectral imaging system, this paper completed the spectral image acquisition of fresh meat. Acquisition experiments show that the hyperspectral image detection system based on acousto-optic tunable filter is goog at reliability and stability.the system with high flexibility and low repetitive error can satisfy the visualization of meat were fast nondestructive testing research needs.