Multisource and multitemporal data in landcover classification tasks: The advantage offered by neural networks
The research adresses the problem, within the Monitoring Agriculture with Remote Sensing (MARS) project, of land cover classification and acreage assessment based on remotely sensed images in case of lack of optical input data due to cloud cover. An alternative strategy, based on the exploitation of multi-source and multi-temporal data by means of a feed-forward neural network (NN) is proposed and discussed. the results reported in the annexe show that NNs not only provide a useful tool for data fusion but also an extremely powerful means for early and reliable acreage assessment.
Bibliographic Reference: Paper presented: IGARSS '97 (SG), August 5-8, 1997
Availability: Available from (1) as Paper EN 40662 ORA
Record Number: 199710868 / Last updated on: 1997-07-08
Original language: en
Available languages: en