The Building Area Recognition in Image Based on Faster-RCNN

Building area recognition technology is a rising up research topic in the field of pattern recognition and it has a wide application prospect in the fields of military, city planning and other fields. This paper combines building area recognition algorithm based on Faster-RCNN and some of the problems in it are discussed respectively and the corresponding solutions are given. Through grabbing a streets cape at Baidu API and real shot in street with SLR camera, building my own image dataset and doing some corresponding training and tests based on Faster-RCNN after images are annotated. The experimental results show that the algorithm above-mentioned can achieve better recognition results. The main work of this article is collecting a large number of image samples and building own image dataset; annotating the images; building and training building recognition model based on Faster-RCNN and combining ZF convolutional neural network to do building area recognition.