Discrete Atomic Transform-Based Lossy Compression of Three-Channel Remote Sensing Images with Quality Control

Abstract

Lossy compression of remote sensing data has found numerous applications. Several requirements are usually imposed on methods and algorithms to be used. A large compression ratio has to be provided, introduced distortions should not lead to sufficient reduction of classification accuracy, compression has to be realized quickly enough, etc. An additional requirement could be to provide privacy of compressed data. In this paper, we show that these requirements can be easily and effectively realized by compression based on discrete atomic transform (DAT). Three-channel remote sensing (RS) images that are part of multispectral data are used as examples. It is demonstrated that the quality of images compressed by DAT can be varied and controlled by setting maximal absolute deviation. This parameter also strictly relates to more traditional metrics as root mean square error (RMSE) and peak signal-to-noise ratio (PSNR) that can be controlled. It is also shown that there are several variants of DAT having different depths. Their performances are compared from different viewpoints, and the recommendations of transform depth are given. Effects of lossy compression on three-channel image classification using the maximum likelihood (ML) approach are studied. It is shown that the total probability of correct classification remains almost the same for a wide range of distortions introduced by lossy compression, although some variations of correct classification probabilities take place for particular classes depending on peculiarities of feature distributions. Experiments are carried out for multispectral Sentinel images of different complexities.

Digital Image Representation by Atomic Functions: The Compression and Protection of Data for Edge Computing in IoT Systems

“…It was shown that atomic wavelets, constructed using , are the most appropriate tool for discrete data processing, in particular, matrices. These functions are applied in discrete atomic compression (DAC) of digital images [ 49 ]. The algorithm DAC belongs to the class of lossy compression algorithms [ 26 ].…”

“…For instance, each element of matrices of 24-bit full-color digital images has three components: red, green, and blue, each of which is an integer from the range 0, 1, …, 255 [ 1 ]. To provide data compressing, quantization of DAT coefficients and further encoding using lossless compression methods are applied in DAC (see Figure 1 ) [ 49 ].…”

“…It is quantizing that makes this algorithm lossy, i.e., a decompressed image does not match the corresponding source. Nevertheless, the quality loss control mechanism, which ensures the desired loss of quality measured by maximum absolute deviation (MAD), root mean square error (RMSE), and/or peak signal-to-noise ratio (PSNR), was recently developed, and the modes of DAC, which can produce compressed images without visual distortions, were obtained [ 49 ]. Furthermore, it was shown that variation of the DAT structure provides a minor variation of DAC efficiency indicators.…”

Data type, compression and interoperability in geographic information formats / Tipos de datos, compresión e interoperabilidad en los formatos de información geográfica

“…Among the forms of compression, it is appropriate to remember the two main strategies: lossy (as we find in most files -not all-in formats such as JPEG, JPEG2000, MrSID or ECW), or lossless (such as those that we can find in files in formats like TIFF, TAR, MMZX, etc). In the case of the former, their use in the world of geographic information has been (e.g., Zabala and Pons, 2011) and is (e.g., Makarichev et al, 2021) constantly reevaluated and reconsidered due to the interest provided by a very high compression versus the risk of an excessive degradation of data. On the other hand, in the case of the later there is, by definition, no data loss and the differences in the compression ratios are not so great, as long as the correct decisions are made.…”

“…Entre las formas de compresión es adecuado recordar las dos grandes estrategias: con pérdida (como encontramos en la mayoría de ficheros -no todos-en formatos como JPEG, JPEG2000, MrSID o ECW), o sin pérdida (como los que podemos encontrar en ficheros en formatos como TIFF, TAR, MMZX, etc). En el caso de los primeros su uso en el mundo de la información geográfica ha sido (e.g., Zabala y Pons, 2011) y es (e.g., Makarichev et al, 2021) reevaluada y reconsiderada constantemente por el interés que proporciona una compresión muy alta ante el riesgo de acabar degradando excesivamente los datos. En cambio, en el caso de los segundos no hay, por definición, pérdida de datos y las diferencias en las razones de compresión no son tan grandes, siempre que se tomen las decisiones correctas.…”