基于 Kinect 传感器的红外场景增强

摘要

针对 Kinect 红外图像噪声大、对比度低等问题，通过对其进行特征分析提出了增强Kinect 红外场景的方法。首先，通过 OCTM线性规划方法提高红外图像对比度；其次，结合 Ki-nect 红外场景的频域特征，找到噪声频率；最后运用频域带阻滤波器和双边滤波相结合的方法对图像进行图像增强，进而达到去除噪声的同时保持边缘细节的效果。为验证方法的有效性和实用性，对不同客观条件下的红外场景进行了实验，并对多组实拍场景进行了主观和客观方面的测试。通过实验和测试，证明此方法能有效增强 Kinect 红外场景，且在去噪同时很好地保留了图像边缘信息。%Aiming at the problems of large noise and low contrast for Kinect infrared images,an enhancement method was proposed for the Kinect infrared scene through analyzing the characteristics.Firstly,the infrared image contrast was enhanced by the OCTMalgorithm.Secondly,combined with the features of frequency domain for the Kinect infra-red scene,the noise frequency band was found.Finally,in order to get rid of noise and preserve edge details of the in-frared scene image,the images were enhanced with the band stop filter and bilateral filter.In order to prove the effec-tiveness and the practicability of the proposed method,the experiments of different infrared scenes were carried out, and multiple sets of real scenes were tested from subjective aspect and objective aspect.Experiments and the test re-sults show that the proposed method is very effective to enhance Kinect infrared scenes.