|
Based on the theory of a digital DPCM, the calculations and discussions on the simulation of the digital DPCM image compression with optical methods are presented, and an incoherent optical system of image information compression is designed. The primary experimental results are shown in this paper to testify the feasibility of this system. 应用数字DPCM信息压缩的理论,对信息量压缩的光学模拟方法进行计算和讨论,设计了一个完全由光学元件构成的图象信息压缩系统,并展示了验证此系统可行性的初步实验结果。 In the field of pattern recognition and image process, moment invariants have been used as feature descriptors in a variety of object identifying and position finding. Moment invariant is a kind of simple repeated algebra calculation, which is invariant to relative to translation, rotation and size of the object, However, due to the recursive nature of the calculations and the limited speed of microprocessor, the moments were not computable in real-time. In this paper, we present new fast moment invariant computations... In the field of pattern recognition and image process, moment invariants have been used as feature descriptors in a variety of object identifying and position finding. Moment invariant is a kind of simple repeated algebra calculation, which is invariant to relative to translation, rotation and size of the object, However, due to the recursive nature of the calculations and the limited speed of microprocessor, the moments were not computable in real-time. In this paper, we present new fast moment invariant computations based on image projection, which are called projection-moment invariants. The difference between our method and the old ones is that projection-moment translates the two-dimensional image information into one-dimensional using projection and greatly reduces the quantity of data and operation. Under present circumstances, projection process can be easily realized in real-time by applying digital hardware, because it is simply composed of additions. So projectionmoments is a new real-time image process programme based on the combination of software and hardware. 矩不变量是一种简单的重复性代数运算,算法简单,而且具有相对于物体平移、旋转和比例不变性,但由于受到矩本身计算量及微处理器速度、存贮空间等限制,矩不变量的计算还很难达到实时实现.本文提出了一种新的、建立在图象投影基础上的二维矩快速算法,称为投影矩不变量.与以往各种矩算法不同的是,投影矩算法利用投影变换,将二维图象信息压缩成一维信息,大大地减少了数据量和运算量,且投影变换只是简单的加法运算,在现有条件下,易于硬件实现,是一种软硬件结合,可望达到实时处理的新方案.计算机仿真结果也验证了该算法的正确性及快速性. Proposing an adaptive image compression scheme with the combination of VQ in DCT domain having two features: (1) The first is basing on the nonstationary of image signal, the DCT is carried on blocks. Then, according to the energy distribution of transformed coefficients to classify the soures,an adaptive scheme is produced. (2) After the soures are classified, the“Rearrangement of subblocks”which we called is an operation of transformed domain. Thus, a good result of VQ is obtained. The experiments show that... Proposing an adaptive image compression scheme with the combination of VQ in DCT domain having two features: (1) The first is basing on the nonstationary of image signal, the DCT is carried on blocks. Then, according to the energy distribution of transformed coefficients to classify the soures,an adaptive scheme is produced. (2) After the soures are classified, the“Rearrangement of subblocks”which we called is an operation of transformed domain. Thus, a good result of VQ is obtained. The experiments show that this scheme has greater advantages in using the nonstationarity of image soures, increasing VQ efficiency, simplifying VQ design and the convenience of systematical realization. 提出了DCT域结合VQ的图象信息压缩方案。有两个特点:1.基于图象信号的非平稳性等特点,分块DCT后,根据变换域系数的能量分布进行信源分类产生自适应方案,2.分类后的信源进行一个我们称之为“子块重排”的变换域操作,从而得到很好的VQ结果。实验表明本文方案在利用图象信源的非平稳性,提高矢量量化器效率,简化矢量量化器设计,以及系统实现的便利性诸方面具有较大的优越性。
|