立体感知质量评价方法与立体视觉特性研究

发布时间：2018-05-27 04:46

本文选题：人眼立体视觉 + 立体相机　；参考：《天津大学》2016年博士论文

【摘要】：近年来,立体图像处理技术迅猛发展,广泛应用于军事、工业、教育和医疗等领域,改变了人们的生活方式,立体图像技术成为了信号处理领域的研究热点问题之一。人们希望获取良好的立体图像质量,但是立体图像从获取到最终的显示阶段,每一阶段都会不可避免的引入噪声,损伤立体感知效果。因此,探究立体感知质量评价方法已成为立体技术领域亟待解决的研究课题之一。立体相机拍摄技术是影响立体图像质量的首要因素,而其他过程噪声的干扰会进一步影响最终的立体感知效果。为保证立体感知效果达到最佳,无论是立体图像获取阶段,还是立体显示阶段都需要进行立体图像质量评价,以填补立体技术领域立体感知效果评价方法的空缺。因此,本文围绕立体图像获取技术、立体图像质量评价方法以及人眼立体视觉特性展开分析和讨论。首先,为解决立体获取技术不佳所引起的立体感知效果损伤问题,保证视觉立体效果的真实性,本文以立体图像的获取技术为突破点,从分析立体相机在不同拍摄条件下的拍摄准则入手,建立了一种综合性的立体相机拍摄质量评价模型。同时,实验结果验证了本文所提模型在评价立体相机拍摄质量方面的有效性。其次,结合人眼视觉特性的两大立体视觉融合处理机制,即单通道融合模型和双通道融合模型,本文提出了两种新的立体图像质量客观评价模型:基于单通道融合模型的立体图像质量评价方法和基于双通道融合模型的立体图像质量评价方法。其优点是两种方法均有效的结合绝对视差图,大大降低了计算复杂度,节省了评价时间。实验证实人眼立体融合特性在立体图像质量评价领域发挥了重大的作用,两种评价方法都能够有效的评价立体图像的质量。相比较而言,由于单通道融合模型研究较为成熟,且结合了人眼立体视觉显著性特性,故基于单通道融合模型建立的评价方法的整体性能优于基于双通道融合模型建立的评价方法。最后,结合人眼立体视觉特性在立体拍摄领域和立体图像质量评价领域的有效应用,本文还就人眼在立体空间上的对比度敏感度视觉特性展开研究,并建立了相应的立体对比度敏感度视觉特性模型,为立体视觉特性在立体技术领域的应用铺平了道路。
[Abstract]:In recent years, stereo image processing technology has been developed rapidly, widely used in military, industrial, educational and medical fields, and has changed people's way of life. Stereo image technology has become one of the hot issues in the field of signal processing. People want to obtain good stereo image quality, but stereo image from the acquisition to the final display stage, each stage will inevitably introduce noise, damage the stereoscopic perception effect. Therefore, exploring the evaluation method of stereoscopic perception quality has become one of the urgent research topics in the field of stereoscopic technology. Stereo camera technology is the primary factor that affects the stereo image quality, while the interference of other process noise will further affect the final stereoscopic knowledge effect. In order to ensure the best effect of stereoscopic perception, it is necessary to evaluate the quality of stereoscopic images in both the acquisition stage and the stereoscopic display stage in order to fill the gaps in stereo perception evaluation methods in the field of stereoscopic technology. Therefore, this paper focuses on stereo image acquisition technology, stereo image quality evaluation method and human stereoscopic vision characteristics analysis and discussion. First of all, in order to solve the problem of stereo effect damage caused by poor stereo acquisition technology and ensure the authenticity of stereo effect, this paper takes the stereo image acquisition technology as the breakthrough point. Based on the analysis of the shooting criteria of stereo camera under different shooting conditions, a comprehensive evaluation model of stereo camera shooting quality is established. At the same time, the experimental results verify the effectiveness of the proposed model in the evaluation of stereo camera quality. Secondly, combining with the human visual characteristics, two stereo vision fusion mechanisms, that is, single-channel fusion model and two-channel fusion model, are proposed. In this paper, two new stereo image quality evaluation models are proposed: one is based on single-channel fusion model and the other is based on two-channel fusion model. The advantage is that the two methods combine the absolute parallax graph effectively, which greatly reduces the computational complexity and saves the evaluation time. Experimental results show that human eye stereo fusion plays an important role in the field of stereo image quality evaluation. Both methods can effectively evaluate the quality of stereo images. By comparison, because of the mature research of single-channel fusion model, and the combination of human stereoscopic vision, Therefore, the overall performance of the evaluation method based on single channel fusion model is better than that based on two channel fusion model. Finally, combined with the effective application of human stereoscopic vision in stereo photography and stereoscopic image quality evaluation, the contrast sensitivity of human eye in stereo space is studied in this paper. The corresponding stereo contrast sensitivity visual characteristic model is established, which paves the way for the application of stereo vision characteristics in stereo technology.
【学位授予单位】：天津大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】