Few-shot Object Recognition Method for Aerial Drones Based on Model Fine-tuning
Aerial drone object recognition is an urgent demand in modern military and aviation fields.Due to the various functions and types of drones at present,it is difficult to collect a large number of new drone samples for training object recognition models.In order to solve this problem,a few-shot object recognition method for aerial drones based on model fine-tuning is proposed.The meth-od is based on a Faster RCNN architecture.Firstly,the Faster R-CNN model is pretrained by using the data of common drones with a large number of labeled samples.Then,the last classification layer of the infrastructure is replaced by the cosine measurement,and the small balanced datasets of new drones and common drones are constructed and jointed to fine-tune the classification layer with a small learning rate.Experimental results show that with the number of labeled samples of 5,10 and 50,the mAP of few-shot object recognition model based on the model fine-tuning is 88.6％,89.2％and 90.8％respectively,which can meet the requirements of few-shot target recognition task of aerial drones,and the proposed method is better than other few-shot target recognition methods.
万方数据
维普
