A spectral reconstruction method, device and electronic equipment

Abstract

The embodiment of the present application discloses a spectrum reconstruction method, device and electronic equipment, the method comprising: acquiring the spectrum of any basic array unit in the imaging area through an image sensor, any basic array unit is composed of N pixels, N is a positive integer; obtain the value of the specified shooting parameter; input the spectrum of the arbitrary basic array unit and the value of the specified shooting parameter into the neural network, and obtain the described arbitrary basic array unit output by the neural network Compensation amount of the corresponding spectrum; superimposing the spectrum of any basic array unit with the compensation amount of the spectrum corresponding thereto to obtain the reconstructed spectrum of any basic array unit. The embodiment of the present application adopts a digital method to reconstruct the spectrum, which is beneficial to reduce the cost. It can not only be used for the reconstruction of the RGB spectrum acquired by the traditional image sensor, but also suitable for the reconstruction of the fine spectrum acquired by the hyperspectral image sensor. It is beneficial to improve the spatial accuracy of the image.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular to a spectrum reconstruction method, device and electronic equipment. Background technique [0002] Due to the dispersion effect of the lens, the refractive index of the lens is different for different wavelengths of light. When the scene is imaged by the lens, light of different wavelengths has different focal planes, and the composite light at any point on the scene will be transmitted to scattered spatial points after passing through the lens. The spatial distribution of the fine spectrum with object characteristics will shift, which will cause the image sensor to misjudge the position and size of the scene, and the generated image will be distorted. [0003] At present, the method of optimizing the lens group is usually used to reduce the image distortion caused by the dispersion, specifically, including: increasing the number of lenses and using lenses of di...

AI Technical Summary

Problems solved by technology

When the image sensor is a hyperspectral image sensor that measures precise spectra, the existing data compensation methods have great limitations. The actual situation that there are many and narrow channels in the spectrum; in addition, it is necessary to calculate multiple bands and consider the effects of longitudinal dispersion and transverse dispersion. The model is complex and the calculation is huge

Images