Recognition and application of palmprint fusion in hyperspectral images

With the development of science and technology and the progress of society, personal identity verification and security verification have attracted more and more attention. As a biometric identification technology, palmprint recognition has been widely used in the field of identity verification and security verification because of its stability and universality. However, traditional palmprint recognition techniques usually only use visible light images, which makes them vulnerable to counterfeiting. In order to solve this problem, the palmprint image recognition technology acquired by hyperspectral imager has been developed.

Hyperspectral images are images taken at different wavelengths. In the field of palmprint recognition, hyperspectral images can provide more information, including skin color, blood vessel distribution, skin texture, etc. By fusing images of different wavelengths, the accuracy and reliability of palmprint recognition can be improved.

In the palmprint fusion recognition of hyperspectral images, the first problem to be solved is how to obtain high quality hyperspectral images. Traditional hyperspectral cameras are expensive and difficult to popularize. Therefore, how to use the existing equipment and technology to obtain high quality hyperspectral image has become the focus of research. One method is to acquire hyperspectral images using multi-frequency light sources and optical filters. Another method is to obtain hyperspectral images using portable devices such as smartphones.

After obtaining high quality hyperspectral images, the next problem to be solved is how to effectively extract palmprint features. Traditional palmprint feature extraction methods are mainly based on visible light images. However, since hyperspectral images contain more information, new feature extraction methods need to be developed. One approach is to use deep learning techniques to extract palm print features. Another method is to extract palmprint features using multiple wavelength information in hyperspectral images.

In the field of palmprint recognition, the commonly used classification algorithms include support vector machine, neural network and decision tree. However, these algorithms have some problems in processing hyperspectral images, such as high computational complexity and unstable classification results. Therefore, new classification algorithms need to be developed. One way is to use deep learning techniques to classify. Another method is to use multiple wavelength information in hyperspectral images for classification.

The palmprint fusion recognition technology of hyperspectral image has a wide application prospect. In terms of personal identity verification, the palmprint fusion recognition technology of hyperspectral images can be used for security verification of bank accounts, electronic payments, e-commerce, etc. In terms of public security, the palmprint fusion recognition technology of hyperspectral images can be used for criminal investigation, immigration management, etc.

In short, palmprint fusion recognition of hyperspectral images is a biometric recognition technology with wide application prospects. The accuracy and reliability of palmprint recognition can be improved by obtaining high quality hyperspectral images, extracting palmprint features and selecting appropriate classification algorithms. With the development of technology and the increasing social demand for security, the palmprint fusion recognition technology of hyperspectral images will play an increasingly important role in the field of identity verification and security verification.