Avaliação da irrigação mecanizada automatizada por câmera RGB em diferentes condições operacionais em plantio de eucalipto

Abstract

The increase in the consumption of forest products has been imposing increasing pressure on the forest plantation sector, more specifically on silviculture, to deliver quality wood at low cost and at ever-increasing volumes. This demand has led the forest sector to a state of alert, since there is a tendency to reduce labor in the field and thus the search for technological innovations becomes crucial for the current moment. A big challenge lies in the mechanization of silvicultural operations. A clear example is irrigation, which has evolved little in recent years and is basically carried out in a semi-mechanized manner. Due to this demand, some technologies have been tested to try to fill this need. Recently, the FMCopling irrigator was launched, which has an RGB camera which makes the remote identification of seedlings and, by means of an automatic algorithm, performs the command for the release of water for irrigation. Thus, this work aimed to evaluate the effectiveness of a mechanized irrigator automated by RGB camera in Eucalyptus plantations under different operational conditions related to the preparation of the area, i.e. with and without planting realignment, and the direction of movement of the irrigator, i.e. towards or away from the sun. As a result, an average of 91.4% was obtained in the identification of seedlings, with no significant difference between the operational conditions tested in relation to the area preparation with or without planting realignment and in relation to the direction of movement of the tractor towards or away from the sun. Regarding the volume of water dispensed per seedling, the average values obtained in the manual and mechanized basin simulations were below the target value for which the implement was calibrated, which indicates the need for an adjustment between the calibration of the implement when still and the volume of water to be dispensed when moving during operation. Finally, mechanized irrigation triggered by identification of seedlings by RGB camera proved to be highly promising.