Using Convolutional Neural Networks to Count Palm Trees in Satellite Images
This work addresses a domain-specific problem for agricultural or environmental monitoring, but it is incremental as it applies existing CNN methods to a new dataset without major methodological innovations.
The paper tackles the problem of counting and localizing palm trees in high-resolution satellite imagery by using a convolutional neural network classifier in a sliding window approach, achieving over 99% tree count accuracy on a small dataset of 500 images.
In this paper we propose a supervised learning system for counting and localizing palm trees in high-resolution, panchromatic satellite imagery (40cm/pixel to 1.5m/pixel). A convolutional neural network classifier trained on a set of palm and no-palm images is applied across a satellite image scene in a sliding window fashion. The resultant confidence map is smoothed with a uniform filter. A non-maximal suppression is applied onto the smoothed confidence map to obtain peaks. Trained with a small dataset of 500 images of size 40x40 cropped from satellite images, the system manages to achieve a tree count accuracy of over 99%.