Presented here is a study comparing various inter-slice bit-allocation strategies when applying JPEG2000 part 2 to ultraspectral sounder satellite data. Two families of algorithms are compared: the first is strictly variance based for which traditional approaches and some variations are used; the second uses rate distortion curves (RDCs) to optimally allocate slice bit-rates in the MSE sense, analogous to the way the PCRD algorithm uses RDCs of code-blocks in the baseline JPEG2000 method. The rate distortion curves are either experimentally gathered or analytically modeled using techniques discussed in previous papers. An additional approach to gather RDCs is considered using the slope-length information computed by a JPEG2000 encoder. The study is done using six different granules of ultraspectral sounder data made available by NOAA. Every subset contains 1,501 spectral bands, each consisting of 90 by 135 pixels, which are decorrelated using either a Karhunen-Loeve transform (KLT) or the discrete wavelet transform (DWT). The most complex RDC-based optimal approach is used to establish a lower distortion bound, in addition to characterizing the typical distortion performance achievable when compressing spectrally decorrelated sounder data with JPEG2000. The bound is used as the basis of comparison against the other methods studied. Moreover, the software tool CompressMD developed by our group is used to provide all data handling, algorithm implementations and collection of results.
This paper describes an efficient algorithm and its Java implementation for a recently developed mean-squared error (MSE) rate-distortion optimal (RDO) inter-slice bit-rate allocation (BRA) scheme applicable to the JPEG2000 Part 2 (J2KP2) framework. Its performance is illustrated on hyperspectral imagery data using the J2KP2 with the Karhunen- Loeve transform (KLT) for decorrelation. The results are contrasted with those obtained using the traditional logvariance based BRA method and with the original RDO algorithm. The implementation has been developed as a Java plug-in to be incorporated into our evolving multi-dimensional data compression software tool denoted CompressMD. The RDO approach to BRA uses discrete rate distortion curves (RDCs) for each slice of transform coefficients. The generation of each point on a RDC requires a full decompression of that slice, therefore, the efficient version minimizes the number of RDC points needed from each slice by using a localized coarse-to-fine approach denoted RDOEfficient. The scheme is illustrated in detail using a subset of 10 bands of hyperspectral imagery data and is contrasted to the original RDO implementation and the traditional (log-variance) method of BRA showing that better results are obtained with the RDO methods. The three schemes are also tested on two hyperspectral imagery data sets with all bands present: the Cuprite radiance data from AVIRIS and a set derived from the Hyperion satellite. The results from the RDO and RDOEfficient are very close to each other in the MSE sense indicating that the adaptive approach can find almost the same BRA solution. Surprisingly, the traditional method also performs very close to the RDO methods, indicating that it is very close to being optimal for these types of data sets.
This paper presents a study on the compression of hyperspectral satellite data using JPEG 2000 and residual encoding (RE). The first step in the process is to apply a decorrelating transform in the spectral-direction or z-direction. In most cases in this study, the Karhunen-Loeve Transform (KLT) is used. For comparison, some examples are also included where the discrete wavelet transform (DWT) is used for this purpose as well as examples with a purely 2-D approach that uses no z-direction transform. Bit-rate allocation techniques are used in order to take advantage of the energy compaction obtained when applying a transform in the z-direction. The transformed slices and their corresponding bit rates are input into JPEG 2000 in order to obtain the compressed bit stream. In this study, the compressed bit stream is decompressed at the encoder side in order to compute the recovered data. These data are then subtracted from the original data in order to calculate the residuals, which are then quantized and losslessly encoded separately using JPEG2000 itself in order to control the maximum absolute error (MAE). An analysis between using and omitting residual encoding with respect to MAE is included. It is observed that a decrease in the MAE by a factor of 3 is achieved for this data with very small overhead when the residual encoding is utilized. The two data sets used in this study are the well known Cuprite radiance imagery from AVIRIS and a set from the Hyperion satellite system, both of which are available in 16 bits per value form.
A bit rate allocation (BRA) strategy is needed to optimally compress three-dimensional (3-D) data on a per-slice basis, treating it as a collection of two-dimensional (2-D) slices/components. This approach is compatible with the framework of JPEG2000 Part 2 which includes the option of pre-processing the slices with a decorrelation transform in the cross-component direction so that slices of transform coefficients are compressed. In this paper, we illustrate the impact of a recently developed inter-slice rate-distortion optimal bit-rate allocation approach that is applicable to this compression system. The approach exploits the MSE optimality of all JPEG2000 bit streams for all slices when each is produced in the quality progressive mode. Each bit stream can be used to produce a rate-distortion curve (RDC) for each slice that is MSE optimal at each bit rate of interest. The inter-slice allocation approach uses all RDCs for all slices to optimally select an overall optimal set of bit rates for all the slices using a constrained optimization procedure. The optimization is conceptually similar to Post-Compression Rate-Distortion optimization that is used within JPEG2000 to optimize bit rates allocated to codeblocks. Results are presented
for two types of data sets: a 3-D computed tomography (CT) medical image, and a 3-D metereological data set derived from a particular modeling program. For comparison purposes, compression results are also illustrated for the traditional log-variance approach and for a uniform allocation strategy. The approach is illustrated using two decorrelation tranforms (the Karhunen Loeve transform, and the discrete wavelet transform) for which the inter-slice allocation scheme has the most impact.