Health Detection on Cattle Compressed Images in Precision Livestock Farming

TL;DR

This paper presents a method for compressing cattle health images in precision livestock farming using Seam Carving and LZW algorithms to reduce energy consumption in data transmission.

Contribution

It introduces a combined image compression approach tailored for livestock health monitoring, optimizing energy use with minimal information loss.

Findings

Achieved a compression rate of 1.82:1 with 13.75s per image

Decompression rate of 1.64:1 with 7.5s per image

Memory usage ranged from 146MB to 504MB

Abstract

The constant population growth brings the needing to make up for food also grows at the same rate. The livestock provides one-third of humans protein base as meat and milk. To improve cattles health and welfare the pastoral farming employs Precision Livestock farming (PLF). This technique implementation brings a challenge to minimize energy consumption due to farmers not having enough energy or devices to transmit large volumes of information at the size are received from their farms monitors. Therefore, in this project, we will design an algorithm to compress and decompress images reducing energy consumption with the less information lost. Initially, the related problems have been read and analyzed to learn about the techniques used in the past and to be updated with the current works. We implemented Seam Carving and LZW algorithms. The compression of all images, around 1000 takes a time of 5 hours 10 min. We got a compression rate of 1.82:1 with 13.75s average time for each file and a decompression rate of 1.64:1 and 7.5 s average time for each file. The memory consumption we obtained was between 146MB and 504 MB and time consumption was between 30,5s for 90MB to 12192s for 24410 MB, it was all files.