|Paper title||UAV-based identification of stubble in forestry plantations|
|Form of presentation||Poster|
Assessing the effects of forest restoration is key to translating advances in restoration science and technology into practice. It is important that forest management learns from the past and adapts restoration strategies and techniques in response to changing socio-economic and environmental conditions (Bautista and Alloza, 2009). However, evaluating restoration over time is a complex task. It requires the measurement of variables that reflect the ecological quality of the systems under restoration in a quantifiable way, so that the process and its changes can be analysed on an objective basis (Ocampo-Melgar et al., 2016). When restoration includes active restoration work, such as planting, monitoring should be based, among other things, on the measurement of attributes of the vegetation planted, as well as the effects of the vegetation on the environment.
One variable measured is the response of the planted vegetation, assessed as survival and growth. This is of interest as it occurs at a rate that makes it possible to distinguish significant changes over short periods of time. On the other hand, it is also interesting to know the response of the introduced vegetation because this vegetation will affect properties of the system under restoration in the longer term. Monitoring this response, albeit in the short term, will make it possible to anticipate the transforming capacity of this vegetation. All this analysis has motivated the development and exploitation of new methods for calculating parameters that analyse the monitoring of a plantation.
In this context, the development of new vegetation monitoring methodologies based on the capture of information with unmanned aerial vehicles (UAV) has become very attractive for improving the characterisation and monitoring of vegetation.
The general objective of this study is the development of an applied technology service for the monitoring of reforestation, characterising the structure of the reforestation, its growth and mortality. The methodology developed involves the planning of data acquisition using RGB (red green blue) and NIR (near infrared) cameras on board low-cost UAV platforms, and the processing of the images obtained.
The study has been carried out in a eucalyptus plantation in Huelva (Andalusia, Spain) where it is necessary to identify the plants in the shortest possible time so that they can be replaced in the months after planting.UAV flight planning was carried out at different months after planting, with both types of cameras, with and without NIR channel, and at different flight heights. The identification of dead trees after 2 months of planting was only possible with cameras incorporating near infrared, and from 4 months onwards at a height of 100m.