Microwave sensors for detection of wild animals during pasture mowing

Abstract

Abstract. More than 400000 wild animals are killed or severely injured every year during spring time pasture mowing. Conventional methods for detection and removal or expulsion of animals before mowing are either inefficient or very time-consuming. The first really working method is based on a pyro-detector which senses the temperature contrast between the animals body and the surrounding pasture. Unfortunately, the detection reliability of this sensor decreases with increasing ambient temperature and strong sunlight, i.e. for typical weather conditions, when pasture is mowed, especially around noon. In this paper, a detector is presented that exhibits complementary behaviour. It works best during dry conditions (i.e. around noon), but has a tendency to false alarms when dew is present (i.e. morning and evening). The sensor is based on a commercial, low-cost Doppler module at 24GHz. It senses the difference of radar cross section between the animals body (high water content, specular reflection) and the pasture (low water content, diffuse reflection). The signal is analysed by means of a non-linear Wigner time-frequency transformation. Experimental results are presented for a laboratory setup as well as for measurement in actual spring-time pasture. The results prove that a microwave sensor is capable of reliably detecting animals of the size of a fawn even if it is covered by a layer of pasture.