实时自适应量化法——逐级均值法

本文以跟踪电视系统中自适应量化器为设计背景,提出了一种新的、实时自适应的快速图象量化方法——逐极均值法,文中首先用Lloyd-Max最佳量化理论分析了这种量化方法的均方误差失真,讨沦了图象中存在孤立亮点时的处理方法。然后论述了这种量化方法应用于跟踪电视系统中的性能,即实现的简单、快速性;对照度变化的自适应性;及图象对比度增强效果。文中通过图象处理实验结果验证了这种量化方法的性能和理论分析的正确性。最后得出结论:逐极均值法量化器是一种能够代替LlodyMax最佳量化器的次佳量化器,这种量化器可以很好地满足跟踪电视系统中对自适应量化器的设计所提出的各方面性能要求;它对那些要求实现简单、实时自适应的量化器应用领域也将具有一定意义。

<div> A new real-time adaptive high-speed image quantization technique, called as step-by-step Mean Approach(SMA), is presented on the background of designing adaptive quantizer in the video tracking system. In this paper, we discuss the performances of this quantizer. The high-speed and the simplicities in implementation, the adaptability to changing illuminance as well as contrast enhancement effect are considered in detail. The mean-squareerror distortion of the quantizer are analyzed by the LLoyd-Max optimum quantization theory.A method to delate the isolated extremely brightening points in image are applied.A Variety of performances of given quantization approach and the results of theoretical analysis are verified through a number of image processing experiments. <span style="font-size: 12px;">Finally, we've got the conclusion: the SMA quantizer is a sub-optimum quantization approach according to Lloyd-Max optimum quantization theory. It can very well meet all perforemance requirements of adaptive quantizer in the video tracking system.</span></div>