Small target detection is a difficult point in target detection. Small target detection needs to identify the location and type of targets with few pixels in the picture and little resolution and feature information, and the algorithms used in the current application of mature medium and large target detection do not work well in detecting small targets. Therefore, improving the capability of small target detection is a current challenge in the field of target detection and an important research direction. In this paper, we will focus on deep learning small target detection technology, first introduce the definition of small targets and the reasons for the difficulty of small target detection, then comprehensively discuss the methods to improve the effectiveness of small target detection, and finally introduce the common small target datasets and the evaluation index of detection algorithms.
Classification, segmentation, and detection are the most important tasks in computer vision, and target detection as one of them is a hot research topic in the field of computer vision, which is widely used in medical, traffic, surveillance, etc. YOLOv4 and R-CNN have excellent target detection performance, and an improved YOLOv4 target detection algorithm is proposed to improve the real-time detection of small targets for target recognition. A priori frames are designed using the K-means clustering algorithm for adapting to different small and medium sizes; a feature layer is extracted according to the size of small and medium-sized labeled objects and four different feature layers are fused for detection; the Mish activation function is applied to the neck of the detection model to improve the detection performance. The experimental results show that the improved algorithm can effectively improve the detection accuracy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.