Research Of Hyperspectral Imaging System Based On Spatial Light Modulator

Since 1980s, the emergence of multi-spectral imaging technology has attracted the attention around the world and started extensive researches. Multi-spectral imaging technology analysis the information of the respective spectral bands of the image, and gradually developed into researching the continuity and relevance of each spectral band, that is hyperspectral imaging technology. Hyperspectral imaging technology enables extracting fine spectral information of each pixel, realizing a better understanding of things, as a result of intensive research and development is widely used in military, industrial and agricultural fields.Hyperspectral imaging system solve the problem of obtaining spectral information which is difficult to the general camera. Although the study of spectral imaging system has been over decades, as the need for the accuracy of spectral resolution and spectral SNR is rising, the cost of the system components, optical performance, the difficulty of data storage capacity and data redundancy, etc. are also increasing. The paper first study based on a prism-mask-based video hyperspectral imaging system, which consists of two optical paths, one way to capture the spectral data of the scene which are under-sampled, one way to capture the complete space information of the scene. Spectral resolution, the luminous flux and spatial resolution are the main index to measure a system’s performance, we put forward the spectral SNR model of the system as a supplement theory of system performance, then through experimental analysis, the spectral SNR of the system proves to be improved as the luminous flux improving.As the system utilizes occlusion mask with fixed sampling point to undersample the incident light, which limiting the research and application in adaptive sampling, the reduction of the amount of data and so on. Based on the existing hyperspectral video imaging system, the article introduces a spatial light modulator to replace the mask, that is hyperspectral imaging system based on SLM, which could sample the incoming light according to the different characteristics of the scene. The article proposes texture adaptive sampling method to reduce the amount of data without affecting the accuracy of the reconstruction of the spectral data, and then proves the method is feasible by experimental analysis.Finally, we built the practical spatial light modulator based hyperspectral video imaging system, after brief introduction, we built a mathematical model for the system, to introduce Compressive Sensing theory into the system.