Effects of Image Compression Techniques on Remote Sensing Images

EFFECTS OF IMAGE COMPRESSION TECHNIQUES ON REMOTE SENSING IMAGES

Dr. Jyoti Sarup1,Arpita Baronia2

1Civil Engineering Department, Maulana Azad National Institute of Technology ,Bhopal,India

2Remote Sensing and GIS, Maulana Azad National Institute of Technology ,Bhopal, India

Email:,

ABSTRACT:

In the past few years, there has been a increase in the need for the amount of information stored in the form of images especially from Remote Sensing Satellites.The volume of digitized image being very high, will considerably slowdownthe transmission and storage of such images. Therefore there is strong need of compression of the images by extracting the visible elements which are encoded and transmitted. This paper compares different image compression techniques such as JPEG ,JPEG2000and SPIHT(Set Partitioning in Hierarchical Tree) using a set of objective picture quality measures like PSNR and Mean Square Error (MSE) and comparison has been done based upon the results of these quality measures. Standard test images were assessed with different compression ratios. It is found that the JPEG2000 based compression has achieved better results as compared to SPIHT and JPEG.

Keywords: Image compression, JPEG, JPEG2000. Lossless compression, lossy compression, Remote sensing, SPIHT.

INTRODUCTION

Image compression can be of two types (1) Lossless and (2) Lossy compressions. With lossless compression, every single bit of data that was originally in the image remains after the image is decompressed. On the other hand, lossy compression reduces an image by permanently eliminating certain information, especially redundant information [1]. In Still images there are many efficient compression techniques. Mainly they are JPEG [2] which is based on discrete cosine transform, SPIHT [3] and JPEG 2000[4] which is based ondiscrete wavelet transform.

JPEG is a popular and continuous tone still image compression mechanism established by first Joint Photographic expert Group in 1992. JPEG is based on Discrete Cosine Transformation of encoder and decoder both. It is a block based technique; where the original image is divided into small n x n (usually 8x8) blocks and then DCT transformation is applied. The data compression is achieved via quantizationfollowed by Huffman coding. The disadvantage of JPEG is the blocking artifacts in reconstructed image [5].Unlike the case of DCT is composed on cosine functions here as DWT can be composed on function (wavelet) which satisfies the multi resolutions. The choice of Wavelet depends on contents and resolution of image in recent time, much of the research activities in image coding have been focused on the Discrete Wavelet Transform (DWT). DWT offers adaptive spatial-frequency resolution (better spatial resolution at highfrequencies andbetter frequency resolution at low frequencies) that is well suited to the properties of Human Visual System (HVS). It can provide better image quality than DCT, especially at higher compression ratio [6].

The SPIHT algorithm was introduced by [3]. It is a powerful, efficient and computationally simple image compression algorithm. By using this algorithm, the highest PSNR values for given compression ratios for a variety of images can be obtained. SPIHT was designed for optimal progressive transmission, as well as for compression. One of the important features of SPIHT is that at any point during the decoding of an image, the quality of the displayed image is the best that can be achieved for the number of bits input by the decoderup to that moment.The wavelet coefficients can be referred as ci,j. The main aim in progressive transmission is to transmit the most important image information at first priority[7]. JPEG 2000 is based on the idea that the coefficients of a transform that decorrelates the pixels of an image can be coded more efficiently than the original pixel themselves. If the transform basis function is wavelet then JPEG 2000 pack most of the visual information into a small number of coefficients, the remaining coefficients can be quantized coarsely or truncated to zero with little image distortion [8].

The following steps are followed for comparisons of compressiontechniques: 1. Select the images (like PANCHROMATICimage). 2. Apply the compression techniques (SPHIT,JPEG and JPEG-2000) on these images. 3. Evaluate the qualityand impact of different technique on image interpretability, finally, a quantitative evaluation of compressed images in orderto estimate the MSE and the PSNR comparisons with theoriginal images.

INDENTATION AND EQUATION

The quality measure for an image is evaluated by MSE andPSNR shown in equation (1) and equation (2).

Mean SquareError (MSE): The mean square error measuresthe error with respect to the center of the image values, i.e.the mean of the pixel values of the image, and by averagingthe sum of squares of the error between the two images.

(1)

where u(x,y) is the original image, v(x,y) is the approximated version (which is actually the decompressed image) and M,N are the dimensions of the images. A lower value of MSE signifies lesser error in the reconstructed image [9].

Peak signal-to-noise Ratio (PSNR): The peak signal-to-noise ratio (PSNR) measures the estimates of the quality of reconstructed image compared with the original image and is a standard way to measure image fidelity. Here signal corresponds to the original image and noise corresponds to the error in reconstructed image due to compression and decompression. The PSNR is a single number that reflects the quality of thereconstructed image and is measured in decibels (db) [10].

Where S is the maximum pixel value and RMSE is the root mean square error of the image. The actual value of the PSNR is not meaningful but the comparison between two values between different reconstructed images gives one measure of quality. As seen from inverse relation between the MSE and PSNR, a low value of MSE/RMSE translates to a higher value of PSNR, thereby signifying that a higher value of PSNR indicates higher reconstruction fidelity [11].

EXPERIMENTAL RESULT AND ANALYSIS

Experiments are conducted on the test image (PANCHROMATIC) using MATLAB platform, coded with JPEG, JPEG 2000, SPIHT image compression coder for each test image with 5 different compression ratios (CR) as 2:1 , 10:1, 12:1,14:1,20:1. TableI shows the MSE of three algorithms between original and decompressed image, Table II Shows PSNR value which is always be greater than the MSE for good result. Fig.1and Fig.2 shows the graphical representation of tables. Fig.3 shows the original image in Tiff Format Fig. 4, 5 and 6 Shows different decompressed images at compression ratio 12:1. The result consists of comparison between three compression technique methods on the basis of calculation of MSE and PSNR of original image and decompressed image.

Table I. MEAN SQUARE ERROR(MSE) Table II Peak Signal To Noise Ratio(PSNR)db

Compression Ratio / SPIHT / JPEG2000 / JPEG
2:1 / 0.70723 / 0.34938 / 6.25778
10:1 / 1.5086 / 1.1506 / 16.0592
12:1 / 0.89324 / 0.50849 / 6.56521
14:1 / 0.66869 / 0.29009 / 6.36022
20:1 / 0.70716 / 0.37423 / 6.36022
Compression Ratio / SPIHT(db) / JPEG2000(db) / JPEG(db)
2:1 / 12.354 / 13.5964 / 11.1657
10:1 / 11.2309 / 12.3603 / 10.1506
12:1 / 12.354 / 13.5964 / 11.1654
14:1 / 12.354 / 13.5964 / 11.1654
20:1 / 12.354 / 13.5964 / 11.1654

Fig. 1. MSE between Original and Decompressed Images

Fig. 2. PSNR between Original and Decompressed Image

Fig. 3. Original Image

Fig. 4 JPEG 2000 Fig. 5. SPIHT Fig. 6. JPEG

CONCLUSION

Based on the limited testing results obtained in this study, it is to be concluded that-There could be a decrease in image quality with compression ratio increase. JPEG2000 has better performance than SPIHT and JPEG. JPEG has poor performance than all the compression methods because all other methods are Wavelet based. Wavelet-based compression provides substantial improvement in picture quality because of overlapping basis functions and better energy compaction property of wavelet transforms.makes imagessmoother and preserves object edges, while DCT-based JPEG creates blocking artifacts. JPEG2000 is better than SPIHT Compression technique. These techniques are scene dependent and for this study area JPEG2000 perform better than the SPIHT.

BIBLIOGRAPHY

Dr. Jyoti Sarup She is designated as a Associate Professor at Maulana Azad National Institute of Technology ,Bhopal (M.P.) India Department of civil Engineering. She has been worked with rolta Technology and Indian Institute of technology, Mumbaias research Associate.Her area of Research is Remote Sensing, Photogrametry,Microwave Remote Sensing, GIS, Image Processing. She completed her P.hd in water resource management.

Arpita Baronia she is designated as a senior Research Fellow at Maulana Azad National Institute of Technology ,Bhopal (M.P.) India Department of civil Engineering. She did her Bachelors Degree in Computer Science and Engineering and M. tech degree in Remote Sensing and GIS from MANIT, Bhopal. Her Area of Interest is Remote Sensing, Image Processing.

REFERENCES

[1] P. J.Paul, S. Koteswari,. and B. K Rani, “A Novel VLSI Architecture of SOC for Image Compression Model for Multimedia Applications”, International Journal of Computer Science and Technology vol. 2, Issue 3, 2011,pp.130-134.

[2] G.K.Wallace, “The JPEG still picture compression standard”, IEEE Transactions on Consumer Electronics, February 1992, Vol. 38,1992, pp. xviii-xxxiv.

[3] A. Said, W. A. Pearlman, "A New, Fast, and Efficient Image Codec Based on Set Partitioning in Hierarchical Trees", IEEE Transactions on Circuits and Systems for Video Technology, Vol. 6, No. 3, 1996, pp. 243-249.

[4] A.N. Skodras, C.A. Christopoulos and T.Ebrahimi,”JPEG2000:The upcoming still image compression method.Pattern Rcognition”, letters 22, 2001, pp.1337-1345.

[5] B. Rani, R.K. Bansal, and S.Bansal, “Comparison of JPEG and SPIHT Image Compression Algorithms using Objective Quality Measures” ,IMPACT’09, pp. 90-93, 2009.

[6] Giby Jose and N. Murugananth,”Image Identification using Compression Technique”, International Conference on VLSI, Communication and Instrumentation (ICVCI), 2011, pp.29-31.

[7] S.P. Raja, P. N.Narayanan, S.Arif Abdul Rahuman,S.Kurshid Jinna and S.P. Princess,”Wavelet Based Image Compression: a Comparative Study”, International Conference on Advances in Computing, Control, and Telecommunication Technologies, 2009, pp. 545- 549.

[8] C. Christopoulos, A. Skodras, and T.Ebrahimi,”The JPEG2000 Still Image Coding System: An Overview”, IEEE Trans. On Consumer Electronics, Vol.46, No.4, 2005,pp. 1103-1127.

[9] I. Avcbas, B. Sankur, and K. Sayood, “Statistical evaluation of image quality measures”, Journal of Electronic Imaging, vol.11, no.2 , 2002,pp. 206 -223.

[10] D.Santa-Cruz, , T.Ebrahimi, J. Askelof, M. Larsson, and C. Christopoulos, “JPEG 2000 still image coding versus other standards”,SPIEs 45th Annual Meeting, Applications of Digital Image Processing XXIII, Proceedings of SPIE, San Diego, California, Vol. 4115, 2000,pp. 446- 454.

[11] R.C.Gonzalez, R.E. Woods, and S.L.Eddins, “ Digital Image Processing Using Matlab”, secondedition (New Delhi:Tata Mc Graw Hill education Private Limited) pp. 374, 2011.

Bibliography Of Authors: