Abstract
In order to solve the problem that, in complex and wide traffic scenes, the accuracy and speed of multi-object detection can hardly be balanced by the existing object detection algorithms that are based on deep learning and big data, we improve the object detection framework SSD (Single Shot Multi-box Detector) and propose a new detection framework AP-SSD (Adaptive Perceive). We design a feature extraction convolution kernel library composed of multi-shape Gabor and color Gabor and then we train and screen the optimal feature extraction convolution kernel to replace the low-level convolution kernel of the original network to improve the detection accuracy. After that, we combine the single image detection framework with convolution long-term and short-term memory networks and by using the Bottle Neck-LSTM memory layer to refine and propagate the feature mapping between frames, we realize the temporal association of network frame-level information, reduce the calculation cost, succeed in tracking and identifying the targets affected by strong interference in video and reduce the missed alarm rate and false alarm rate by adding an adaptive threshold strategy. Moreover, we design a dynamic region amplification network framework to improve the detection and recognition accuracy of low-resolution small objects. Therefore, experiments on the improved AP-SSD show that this new algorithm can achieve better detection results when small objects, multiple objects, cluttered background and large-area occlusion are involved, thus ensuring this algorithm a good engineering application prospect.
Highlights
Pedestrian and vehicle object detection and recognition in traffic scenes is an important branch of object detection technology and the core technology in the research fields of automatic driving, robot and intelligent video surveillance, both of which highlights its significance in research [1].The object detection algorithm based on deep learning can be applied to a variety of detection scenarios [2], mainly because of its strong comprehensiveness, activeness and capability of detecting and identifying multiple types of objects simultaneously [3,4,5,6]
In order to solve the problem that the existing SSD is difficult to detect small and weak objects with low resolution in complex large scenes, this paper proposes a dynamic region zoom-in network (DRZN), which reduces the calculation of object detection by down-sampling the images of high resolution large scenes while maintaining the detection accuracy of small and weak objects with low resolution in high resolution images through dynamic region zoom and the effect of improving the detection and recognition accuracy is obvious
In order to solve the problem that in complex large traffic scenes, we can hardly balance between the accuracy and real-time performance when we use existing object detection algorithms based on big data and depth learning, this paper improves the object detection framework SSD based on depth learning and proposes a new multi-object detection framework AP-SSD, which is dedicated to multi-object detection in complex large traffic scenes
Summary
The object detection algorithm based on deep learning can be applied to a variety of detection scenarios [2], mainly because of its strong comprehensiveness, activeness and capability of detecting and identifying multiple types of objects simultaneously [3,4,5,6]. Among various types of artificial neural network structures, deep convolutional networks, with powerful feature extraction capabilities, have achieved satisfactory results in visual tasks such as image recognition, image segmentation, object detection and scene classification [7]. Faster R-CNN (where R corresponds to “Region”) [8] is the best method based on deep learning. The RPN [8] network and the Fast R-CNN network are combined and the proposal acquired by the RPN is directly connected
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.