Abstract
Most traditional object detection approaches have a deficiency of features, slow detection speed, and high false-alarm rate. To solve these problems, we propose a multi-perspective convolutional neural network (Multi-PerNet) to extract remote sensing imagery features. Regions with CNN features (R-CNN) is a milestone in applying CNN method to object detection. With the help of the great feature extraction and classification performance of CNN, the transformation of object detection problem is realized by the Region Proposal method. Multi-PerNet trains a vehicle object detection model in remote sensing imagery based on Faster R-CNN. During model training, sample images and the labels are inputs, and the output is a detection model. First, Multi-PerNet extracts the feature map. Meanwhile, the area distribution and object-area aspect ratio in the sample images are obtained by k-means clustering. Then, the Faster R-CNN region proposal network generates the candidate windows based on the k-means clustering results. Features of candidate windows can be obtained by mapping candidate windows to the feature map. Finally, the candidate window and its features are inputted to the classifier to be trained to obtain the detection model. Experiment results show that the Multi-PerNet model detection accuracy is improved by 10.1% compared with the model obtained by ZF-net and 1.6% compared with the model obtained by PVANet. Moreover, the model size is reduced by 21.3%.
Highlights
Remote sensing object detection has been a widely pursued research topic in the field of remote sensing
The third is based on statistical theory. It employs machine learning, the support vector machine (SVM), and other methods, such as random forest (RF) [2] and the backpropagation (BP) neural network [3]. This method is mainly to extract the feature of the target by RF, backpropagation artificial neural network (BPANN), et al using the feature, the classification model can be trained by SVM
The statistical theory approach is used for target detection
Summary
Remote sensing object detection has been a widely pursued research topic in the field of remote sensing. We consider the image features extracted from convolutional networks with different initial receptive fields, and we reduce the influence of the number of samples on the detection accuracy by increasing the extracted image features. Geo-Inf. 2018, 7, 249 the influence of the number of samples on the detection accuracy by increasing the extracted image features. The sTechoenbdacsoicnsvtorulucttiuorneaflolrayeearchisfedaotwurnesaemxtpralectdioanndnetthweorreksuisltsshoofwthneinsaFmigpulirneg7.arOencoemacbhinneedtwwoirtkh, tthhee sfeecaotunrdescoenxvtroalcutteidonbayl ltahyeelraisst dcoonwvnoslaumtipolneadl alanyderthteo roebstualitns oFfeatht_emsaamp,palisnfgolalroewcso:mbined with the features extracted by the last convolutional layer to obtain Feat_map, as follows: Feat_mapi = down_conv2 + convj Here, i is branch i, and j is coFnevaotl_umtiaopnia=l ladyoewrnj._conv2 + convj. The second convolutional layer is downsampled and the results of the sampling are combined with the features extracted by the last convolutional layer to obtain Feat_map, as follows: Feat_mapi = down_conv2 + convj. For Multi—PerNeti, i represents the number of the basic structure of the feature extraction network
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
