CCSDS120.1-G-1：国际空间数据系统图像压缩标准

需积分: 50 107 浏览量更新于2024-07-25 1 收藏 5.44MB PDF 举报

"CCSDS120.1-G-1是关于空间数据系统标准的一份报告，重点关注图像数据压缩的国际标准。这份英文版报告由咨询委员会为空间数据系统（CCSDS）的管理委员会批准发布，反映了CCSDS成员国机构技术工作组专家的共识。报告的审查和授权流程遵循CCSDS的程序手册。该文档由CCSDS秘书处发布并维护，位于美国华盛顿特区的国家航空航天局办公室。" CCSDS（Consultative Committee for Space Data Systems，空间数据系统咨询委员会）是一个国际组织，致力于制定和推广太空通信与数据处理的标准。在图像数据压缩领域，CCSDS制定的标准对于确保不同国家和机构之间的空间任务能够有效交换和处理图像信息至关重要。报告"GREENBOOK, Issue 1"是2007年6月发布的第一版，其核心内容可能涉及了多种图像压缩算法和技术，旨在优化在有限带宽和存储资源下的空间通信效率。这可能包括了有损和无损压缩方法，如预测编码、变换编码（如DCT - 离散余弦变换）、熵编码（如霍夫曼编码和算术编码）、以及适应性编码策略等。报告可能会详细讨论针对空间环境的独特挑战，比如高辐射环境对硬件的影响、长时间延迟的通信问题、以及对数据恢复和图像质量的严格要求。此外，考虑到空间任务的多样性，报告可能还涵盖了不同类型的图像数据（如多光谱、高分辨率地球观测图像、星际探测图像等）的特定压缩需求。在标准制定过程中，通常会进行详尽的性能评估，包括压缩比、解压速度、图像质量（通过PSNR - 峰值信噪比和SSIM - 结构相似性指标等度量）等方面的测试。这些评估确保了压缩标准在实际应用中的有效性。 CCSDS120.1-G-1报告是国际空间图像压缩领域的关键参考资料，对于理解和实施太空通信中的高效数据压缩技术具有重要意义。这份2007年的绿皮书可能已经为后续的空间任务和数据处理技术奠定了基础，并且随着时间的推移，这些标准可能会不断演进和完善，以适应技术进步和新出现的挑战。

CCSDS REPORT CONCERNING IMAGE DATA COMPRESSION

segment boundaries. Consequently, users of the Recommendation must employ a suitable

packetization scheme (see reference [2]) or other means of identifying segment boundaries.

Within each compressed image segment, data are arranged so that earlier portions of the

compressed data segment tend to make a larger contribution in overall reconstructed segment

fidelity than later portions. This embedded data structure allows a user to meet a constraint

on compressed segment data volume by truncating the compressed bitstream for a segment at

the appropriate point.

The tradeoff between reconstructed image quality and compressed data volume for each

segment is controlled by specifying the maximum number of bytes in each compressed

segment along with a ‘quality’ limit (see 3.4). The quality limit constrains the amount of

wavelet-transformed image information to be encoded for each segment. For each segment,

compressed data is produced until the byte limit or quality limit is reached, whichever comes

first. Lossless compression is achieved when the integer DWT is used and the number of

bytes required for losslessly encoding each segment does not exceed the segment byte limit.

2.2 QUANTIFYING COMPRESSION EFFECTIVENESS

For a compressed image, the bit rate achieved by the compressor, measured in bits/pixel, is

defined as the number of bits used in the compressed representation of the image divided by

the number of pixels in the image.

When image compression is lossy, one of several distortion or quality metrics can be applied

to quantify the degree to which the reconstructed image matches the original. Examples of

distortion or quality metrics include:

– Mean Squared Error (MSE):

w ⋅ h

i, j

−

i, j

)

∑

–

Peak Signal to Noise Ratio

(PSNR), measured in dB:

20log

−1

MSE

(dB)

–

Maximum Absolute Error (MAE)

max

i, j

−

i, j

Note that other definitions of PSNR have been used in the literature, and so one should be cautious when

comparing PSNR values from two different sources.

CCSDS 120.1-G-1 Page 2-2 June 2007

CCSDS REPORT CONCERNING IMAGE DATA COMPRESSION

Here and h denote the image width and height, respectively; w

i, j

and

i, j

denote the

original and reconstructed pixel value in the

row and

column of the image; and

i j

denotes the dynamic range (in bits) of the original image.

Technically speaking, PSNR is a quality metric (because increasing PSNR values indicates

increasing reconstructed image fidelity) while MSE and MAE are distortion metrics.

Note that when compression is lossy, the reconstructed image (and therefore, image

distortion) depends on the choice of scheme used by the decompressor to reconstruct the

image given the information encoded in the compressed image segments. Subsection 4.4

describes the particular scheme used to produce all results presented in this Report.

For a given compressor with a fixed set of parameters, the term rate-distortion performance

or compression effectiveness is used to refer to the image quality achieved as a function of bit

rate. Section 3 includes some examples of rate-distortion curves. In the case of lossless

compression, compression effectiveness is simply measured by the bit rate achieved. Note

that compression effectiveness does not take into account any measure of implementation

complexity (speed, memory requirements, etc.).

2.3 OTHER COMPRESSION APPROACHES

An alternative lossless image compression technique is the JPEG-LS standard (reference [3]).

Compared to the CCSDS Image Data Compression Recommended Standard, JPEG-LS has

lower complexity and achieves similar compression effectiveness, but it is designed only for

lossless and near-lossless compression. Another lossless compression technique, which has

lower complexity but is not specifically tailored for imagery, has been previously adopted by

CCSDS in 1997 (see reference [4]).

The JPEG2000 standard (reference [5]) is another wavelet-based image compressor that is

capable of providing effective lossless and lossy compression. The present Recommendation

differs from the JPEG2000 standard in several respects:

it specifically targets use aboard spacecraft;

a careful trade-off has been performed between compression performance and

complexity;

being less complex, it can be more easily implemented in either hardware or

software;

it has a limited set of options, supporting its successful application without in-depth

algorithm knowledge.

Compression effectiveness comparisons with other compressors are provided in section 5.

CCSDS 120.1-G-1 Page 2-3 June 2007

CCSDS REPORT CONCERNING IMAGE DATA COMPRESSION

2.4 ALGORITHM OPERATION

2.4.1 GENERAL

The compressor consists of two functional parts, depicted in figure 2-1: a DWT module

which performs the DWT decomposition of image data, as described in 2.4.3 of this

document and section 3 of reference [1], and a Bit-Plane Encoder (BPE) which encodes the

transformed data, as described in 2.4.4 of this document and section 4 of reference [1]. In the

remainder of this document it is assumed that the reader is familiar with the contents of

reference [1] including terminology defined there.

Discrete

Wavelet

Transform

Input

Data

Coded

Data

Bit-Plane

Encoder

Figure 2-1: General Schematic of the Coder

2.4.2 INPUT IMAGE

For the purposes of the Recommendation, an image refers to a two-dimensional rectangular

array of integer pixel values. The Recommendation supports compression of two-

dimensional (grayscale) images with pixel dynamic range (bit depth) up to 16 bits per pixel.

Pixels may be signed or unsigned integer quantities.

The Recommendation supports image widths as small as 17 columns and as large as 2

columns. The minimum image height is 17 rows. The maximum image height is unbounded

because it is not explicitly encoded as part of header information, but rather image height is

inferred indirectly. Specifically, Part 1 of the segment header flags the last segment in an

image, which determines the image height to within eight rows. The height can then be

completely determined because the value of the height modulo 8 is encoded (in the PadRows

field) in Part 1 of the segment header for the last segment (see subsection 4.2 of

reference [1]). Because of this indirect method of indicating height, in a push-broom

imaging application one could produce and transmit many compressed image segments

before the height of the image is known.

Values of basic quantities describing the input image (PixelBitDepth, SignedPixels,

ImageWidth) are encoded in the optional Part 4 of the segment header (see 2.5.2). This

header part also indicates whether the image should be transposed (TransposeImg) following

decompression.

CCSDS 120.1-G-1 Page 2-4 June 2007

CCSDS REPORT CONCERNING IMAGE DATA COMPRESSION

2.4.3 DISCRETE WAVELET TRANSFORM

2.4.3.1 Architecture

A program and data flow diagram of the transform (DWT) stage of the compressor is shown

in figure 2-2

2-2

. Image data input is shown at left in the diagram. The DWT coefficient buffer

at right in the diagram stores wavelet coefficients computed by the DWT stage. The program

flow in the diagram produces DWT coefficients for a single segment. Once the coefficients

corresponding to a segment have been computed and placed in the buffer, the BPE stage can

begin encoding that segment. The BPE stage relies on all of the coefficients in a segment

being available simultaneously.

The different levels of DWT decomposition in figure may operate at different clock

rates: the clock rate requirement decreases as the level number increases because of the

reduction of resolution at which higher levels of the DWT process the image.

The calculation of a DWT coefficient depends on image pixels in a limited neighborhood.

Consequently, when images are encoded using more than one segment, it is often possible to

implement the DWT stage so that DWT coefficients for a segment are produced without

requiring input of a complete image frame. For example, in a push-broom imaging

application, DWT coefficients may be produced in a pipeline fashion as new image scanlines

are produced by the imager (references [6] and [7]). A pipeline implementation of the DWT,

as discussed in 4.3, leads to DWT-internal storage requirements that are significantly lower

than for a conventional frame-based implementation. Subsection 4.3 provides more detail on

the program flow within each level and on the size requirements of the DWT-internal buffers

(not illustrated in figure ).

performDWT level 1

performDWT level 2

performDWT level 3

segment

complete?

HL1, LH1, HH1

LL1

yes

HL2, LH2, HH2

HL3, LH3, HH3, LL3

LL2

start

image lines

encode segment

DWT

coefficient

buffer

NOTE – Line-type arrows indicate program flow; block-type arrows indicate data flow.

Figure 2-2: Program and Data Flow of DWT Module

CCSDS 120.1-G-1 Page 2-5 June 2007

剩余117页未读，继续阅读

流年_恍若初见

粉丝: 1
资源: 2

CCSDS120.1-G-1：国际空间数据系统图像压缩标准

CCSDS图像压缩算法研究与FPGA实现及DSP移植

"深空探测与近地观测中的CCSDS图像压缩算法研究与FPGA实现

CCSDS空间数据无损压缩推荐标准

CCSDS图像压缩算

CCSDS图像压缩算法

基于ccsds标准的无损数据压缩算法

空间数据系统咨询委员会(CCSDS)的数据压缩标准文档

空间图像CCSDS压缩算法研究与FPGA实现

2012年CCSDS标准

CCSDS深空通信协议标准

最新资源