The use of multispectral cameras to remotely sense land cover over wide areas has been used for a number of years. More recently the use of multiple wavelength lidar sensors are being studied with the hope of more accurately classifying land cover types.
Researchers at Ryerson University in Canada are reporting impressive results from their use of multispectral lidar sensors that are capable of recording a diversity of spectral reflectance from objects.
They used two different methods to develop land cover classification of an urban study area. The first was image-based classification, where intensity and height images are created from LiDAR points and then a maximum likelihood classifier is applied.
The second is point-based classification, where ground filtering and Normalized Difference Vegetation Indices (NDVIs) computation are conducted. A dataset of an urban area located in Oshawa, Ontario, Canada, is classified into four classes: buildings, trees, roads and grass.
An overall accuracy of up to 89.9% and 92.7% is achieved from image classification and 3D point classification, respectively.