An Efficient Image Compression Method for LoRa Networks
Year: 2026
Type: conference-paper
Abstract: The significant advantages of LoRa technology for use in Internet of Things systems are long range, low power consumption, and high noise immunity. The relatively low data transfer rate typical of Low-Power Wide-Area Network (LPWAN) networks can be considered as disadvantage of LoRa when transmitting images. Typically, low-resolution images are transmitted in such networks. To transmit higher-quality images over a LoRa network, we propose an image preprocessing method that uses the discrete Haar wavelet transform (DHWT) and JPEG compression. The proposed method reduces the file size of the processed image by 1.4 to 2.9 times compared to JPEG compression. The peak signal-to-noise ratio (PSNR) of the restored image ranged from 25 dB to 40 dB, depending on the number of the wavelet transform iterations and the quality level of JPEG compression. Image capture and processing were performed on the ESP32-CAM board. A larger number of iterations takes more time and leads to artifacts due to the accumulation of errors in the wavelet transform process. As a result, by changing the compression quality level and the number of the wavelet transform iterations, it is possible to ensure sufficiently high-quality image transmission over the LoRa network.
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204548818
