Implementación de un sistema de detección automático para repeler aves que se aproximan a los cultivos mediante ondas sonoras utilizando técnicas de visión artificial.

Abstract

The objective of this curricular integration work was to implement an automatic system to repel birds by means of sound waves using artificial vision. For which, it was implemented a software that allows detecting the movement of birds in the crops, aimed to work together with electronic devices. A system of automatic capture of images was implemented which allows to obtain the input of the system, analyzing the images with artificial vision tools oriented to machine learning for the bird detection with the help of OpenCV libraries and Haar cascade classifiers run on free software Linux. A detector algorithm was designed which was programmed on a low-cost embedded card. The system detects the presence of birds to repel them by emitting an audible alarm, which does not abruptly influence on the auditory sense and balance of birds permanently. Through tests it was determined that there are sound waves with a frequency of 44 kHz and a sound intensity of 90 dB, which can achieve the purpose of repelling birds with natural events (such as thunder). The total time for detecting the operation of the system is 7 seconds and the sound emission time is 90 seconds.