Collaborative image compression with error bounds in wireless sensor networks for crop monitoring

Using the correlation characteristic between the reference image may reduce transmission energy in wireless sensor networks. However, as crop images are usually captured periodically and the images are easily influenced by the peripheral environment during capture, transmitting the subtraction information of crop images may consume large amounts of energy. Moreover, sensor nodes should provide sufficiently accurate images for evaluating crop status according to the crop conditions. A compression scheme should be designed to compress the subtraction information and to ensure reconstruction image quality. In this paper, based on the properties of non-standard Haar transformation, we apply non-standard Haar transformation to decompose the subtraction crop images, and we use the error threshold method to compress crop images and to ensure image quality. Experimental results show that our scheme has a higher compression ratio and higher computing efficiency than the Haar wavelets method and the JPEG method.

► Crop images are easily influenced by peripheral environment. ► The subtraction crop images should be compressed to reduce transmission power. ► The Haar wavelet transform can compress crop images in different precision degree. ► Decomposing the subtraction crop images by the Haar wavelet transform has high compression ratio and high computation efficiency.