农业图像感兴趣区域可调压缩处理系统设计与实现

Design and Implementation of a Software System on Compressing Region of Interest of Agricultural Images

随着科技水平的快速发展,图像处理技术也愈加成熟。目前,在基于感兴趣区域的图像压缩对医学图像、遥感图像方面进行较多的研究,但是在农业图像方面的感兴趣分割以及压缩算法研究较少。设计一个压缩比可调的农业图像感兴趣区域压缩处理系统,首先采用支持向量机SVM的模型,对农业图像进行感兴趣区域(ROI)与非感兴趣区域(RONI)的分割处理;再对感兴趣区域采用可调压缩比的JPEG图像压缩算法进行压缩处理,对非感兴趣区域采用高压缩比的JPEG图像压缩算法进行压缩处理,然后将两个区域进行结合,形成完整的农业图像。最后,将处理后的农业图像进行客户端/服务器模式(C/S模式)传输,实现远程传输图像的功能。实验展示,实现4.7:1至5.75:1之间的感兴趣区域(ROI)可调压缩比功能,并能在2秒内实现所有处理,提高传输过程中的传输速率,具有一定应用前景。

With the rapid development of science and technology, image processing technology becomes increasingly more mature. Currently, there are lots of researches on the compression of medical images and remote sensing image based on region of interest, but similar researches on agriculture image segmentation and compression are relatively few. Presents a system on scalable compressing region of interest of agri- cultural images. Segments the agriculture images into region of interest （ROI） and non-ROI （RONI）, uses the support vector machine SVM model; then, applies JPEG image compression algorithm on the ROI with a scalable compression ratio and RONI with a high scalable compression ratio, and combines these two regions into an integral agricultural image; transmits the agricultural image with a C/S model, realizes the remote transmission of images. In our experiments, realizes the compression of ROI with a scalable compression ratio ranging from 4.7：1 to 5.75：1, and this technology allows the whole process to be finished within 2 seconds, which improves the transmission rate and thus has a broad application prospect.