Evolutionary feature selection to estimate forest stand variables using LiDAR

摘要: Abstract Light detection and ranging (LiDAR) has become an important tool in forestry. LiDAR-derived models are mostly developed by means of multiple linear regression (MLR) after stepwise selection predictors. An increasing interest machine learning evolutionary computation recently arisen to improve use LiDAR data processing. Although already proven be suitable for regression, may also applied parametric such as MLR. This paper provides a hybrid approach based on joint MLR novel genetic algorithm the estimation main forest stand variables. We show comparison between our other common methods selecting The results obtained from several datasets with different pulse densities two areas Iberian Peninsula indicate that algorithms perform better than statistically. Preliminary studies suggest lack conditions field possible misuse tests reasons performance algorithm. research confirms findings previous outline importance context analisys data, especially when size fieldwork datatasets is reduced.