Progressive image mosaicking in wireless image sensor networks

Abstract

Prevailing image mosaicking algorithms generate a complete image that incorporates a variety of different images. This variety is captured by different camera sensors. However, these traditional approaches cannot be directly applied to the emerging Wireless Image Sensor Networks (WISNs), because the slow image transmission over a wireless channel causes a noticeable delay before an entire image can be received by its receiver node in a WISN. In this thesis, we propose a Progressive Image Mosaicking Algorithm (PIMA) based on the multi-scan feature of Progressive JPEG (P-JPEG). PIMA's distinguishing characteristic is that it successfully performs mosaicking by using segmental data of images, as opposed to traditional methods, which require the complete data from all images. PIMA mosaics images that are decoded from P-JPEG scans at three levels of quality, and delivers an approximate view of the scene in a short time while the reception of further image data is in progress. Thereafter, it updates the image registration on two other refined levels to gradually enhance the display quality. We developed the concept of Richer Information and Likeliest (RIL) block pair, which is a variation of the Sum of Absolute Difference (SAD) and greatly improves the accuracy of image registration. Experimental results show that PIMA decreases the time delay before the first display of the scene, while preserving an equivalent performance of existing patch-based image mosaicking algorithms.