Original Image (i) 512*64 Dr. Mohammad V. Malakooti (1) , Mehrzad Khederzadeh(2), Amir Tavakoli Golpaygani(3), Dr. Hamidreza Naji(4) (1)Faculty and Head of Department of Computer Engineering, Islamic Azad University ,UAE Branch, Dubai, UAE malakooti@iau.ae (3,4)Graduate Students and (4)Faculty of Department of Computer Engineering, Islamic Azad University ,UAE Branch, Dubai, UAE mz_khederzadeh@yahoo.com , amir.tavakoli.g@gmail.com , hamidnaji@ieee.org Abstract – This paper presents a robust and optimized data compression and suppression based on DCT and Image overlapping methods to transfer panorama images through the Wireless Sensor Networks (WSN). First, we will receive the input images in a fixed interval of times and then we apply two compression algorithms to achieve a compressed and transferable panorama images taken from eight different directions. Our Proposed Algorithm can compress the size of original panorama images up to thirty percent while the vital color information as well as the quality of the original images can be preserved .The panorama image quality and the size of compressed image are two main issues that need to be considered during the process of the panorama image compression. Our DCT-Overlapped Panorama Image compressing Algorithm consists of two major compressing levels of image compression and Overlapping using DCT and our new Overlapping techniques. Keywords:DCT, lossyCompression, image diffusion, panorama, WSN I. INTRODUCTION There are two classes of data compression techniques. The first one is lossy compression and second one is lossless compression. In lossy we will lose some details of image information but still preserved the vital information of the image and the quality of image is in acceptable range. The main goal of lossy compression is to remove redundancy and some details that are not vital to the image and reduce the amount of data storage while preserving the quality of the image. The data compression techniques often are applied on the sound and images that have some redundancy that can be removed to reduce the size of audio and video information before transmission on any networks. A good example of lossy image compression is JPEG [1]. In the second class of compression, lossless compression, the detail of information is important and we would like to preserve all details and vital information while reducing the size of image or sound files for the storage reduction. This type of data compression is more suitable for text and file compression [3]. There are many applications that use lossy data compression techniques to reduce the data redundancy and improve their efficiency by reducing data storage and processing time, such as air and water quality monitoring, Natural disaster prevention, industrial and agriculture monitoring. [2]In this paper, we have presented a robust algorithm based on the DCT and overlap compression for panorama images, in which every image that is received from a WSN has a phase shift of 45 degree respect to the previous image. Then, DCT compression is applied on all images and overlap compression also is applied on the compressed images to reduce the size of panorama images based on the received images. Thus, we are able to reduce the size of panorama image about 70 percent of its original size while preserving the quality of original panorama image. In this process every images is taken with an offset of 45 degree to have a 360 degree panorama image for the monitoring of environment activities. We use our algorithm to reduce the size of panorama image before transmission through WSN node to enhance the energy consumption and prolong the life time of sensors. Figure 1 shows the block diagram compression process. In this figure we can see two levels of compression which are DCT and overlap compressions. In the first level of compression we retrieved the original image by a MATLAB command called imread and then changed it into RGB Color spaces with three different matrixes (Red, Green and Blue). Each matrix is broken down into a set of 8 × 8 blocks, and then the Discrete Cosine Transform (DCT) is applied on each block. The DCT will transfer the original image from the spatial domain into the frequency domain in which the vital information is moved to the low frequency area, Low-Low LL, and less important information, or details, will be moved into Low-High, LH, High-Low HL, and High-High, HH, frequency areas. Once the DCT is applied and the original image is transferred into DCT coefficient matrix of block size 8 x 8, then the last row and last column of the each 8*8 blocks are deleted to obtain a good compression rate while the vital information is preserved and the quality of compressed image is guaranteed. The result of above DCT compression will reduce the size 8 x 8 sub- blocks into 7x7 sub-blocks, and 15 elements of the 8 x 8 blocks are deleted to reduce the size of each block from 8 x 8 into 7 x 7.We will continue this operation until all sub-blocks of 8 x 8 are reduced into sub-blocks of size 7 X 7. This operation will be applied for the sub- blocks of Read, Green, and Blue matrices. Figure 3 shows the block diagram of the DCT Compression Processes. In the second level of compression process the overlapped part of all images will be deled to reduce the size of the underline panorama images combined from all captured images during the 360 camera rotation as shown in Figure 2. In a 360 degree panorama image we have captured 8 images, one at each 45ᴼ rotation. Among these eight images four of them have an overlap cross section of about 23.7ᴼ. In order to have a compressed panorama image we should remove these sections. Figure 2: Panorama image with central position for 360 degree image Our image overlapping algorithm consists of two steps. In the first step of compression the first image, No. 1, is taken by the camera located at the fix position, then the second one, No. 2, is obtain after a 45ᴼ camera rotation, and all other images, No. 3-8, are taken by 45ᴼ camera rotation. Once all images are captured by the camera rotation the DCT is applied on all eight images but the overlapping compression algorithm is applied only on four images, No. 1,3,5,7. In the DCT compression level each image will be divided into blocks of size 8 x 8 and DCT algorithm is applied on each block and then 15 elements of high-frequencies area are removed. Thus the algorithm A Novel Data Compression and Suppression Method based on DCT and Overlapping for the Wireless Sensor Networks(WSN) Separate R, G, B DCT Compression Red, Green, Blue Blocks Copy to Compressed image (J) Overlap Compression Red, Green, Blue Blocks Compressed Image i (448*56) 512*64 448*56 j=j+1 Next Original Image (i) i=i+1, i=1 to 8 Figure1: Block diagram of panorama compression 52 ISBN: 978-0-9853483-3-5 ©2013 SDIWC