Research And Implementation Of Object Recognition Method Based On Deep Neural Network

With the rapid improvement of computer performance,long-standing deep learning algorithms have finally gone in a period of rapid development.Object recognition(also called object detection,target detection)is one of the most valuable research directions in the field of computer vision.This thesis focuses on the application of convolutional neural network algorithm in object recognition in general scenes.More specifically,object recognition here refers to the identification of traffic information(including pedestrians,passing vehicles,traffic lights,etc.)when the vehicle is traveling.The traditional target detection method is divided into three steps.Firstly,the target suggestion boxes are generated on the original image,then the features of these suggestion boxes are extracted,and finally the object in the frame are classified and the frame is adjusted.Each step has its own problem.Target suggestion box generation strategy directly affects the detection speed,accuracy,and computational redundancy.The traditional method of manually extracting image features does not guarantee the quality of the feature.Classification using traditional machine learning methods results into low speed.More importantly,these three steps are completely separate and cannot be used for real-time detection.This thesis aims at the above three problems,using neural network algorithm to break through the difficulties of the problem.Firstly,for the difficulty of manually extracting image features,the thesis implements a convolutional neural network based on dense connection network(DenseNet),which can automatically extract high-quality depth features and can replace the manual extraction of image features.Secondly,for the problem that the speed of traditional classifier is slow,the thesis uses the Softmax classifier to make predictions.The classifier can be used naturally in combination with a convolutional network,so that the second and third steps can be merged into one network,greatly improving the detection speed and accuracy.Thirdly,the thesis abandoned the way to generate the suggestion frame directly on the original image,but use neural network to extract the image feature first,and then propose the strategy on the feature map,which is accurate and efficient.Finally,the thesis applies the above solution to the SSD detection method to generate an improved SSD detection algorithm.After testing,the improved SSD method has significantly improved detection speed and accuracy,and performs better when detecting small targets.The improved SSD detection method incorporates the three detection processes into the same network and truly enables end-to-end real-time detection.At last,this thesis implements a real-time object recognition(identification of traffic information)system based on the improved SSD detection algorithm.Thanks to the advantages of fast detection speed and high precision,the system finally achieves the goal of stable real-time detection.