Abnormal target detection for key components of locomotive based on image processing

2019 ◽  
Author(s):  
Hui Yin ◽  
Jianping Peng ◽  
Wenwei Song ◽  
Xiaorong Gao ◽  
Jianqiang Guo ◽  
...  
Keyword(s):  
Image Processing ◽  
Target Detection

scholarly journals Automatic Acoustic Target Detecting and Tracking on the Azimuth Recording Diagram with Image Processing Methods

Sensors ◽  
2019 ◽  
Vol 19 (24) ◽  
pp. 5391
Author(s):  
Fan Yin ◽  
Chao Li ◽  
Haibin Wang ◽  
Fan Yang
Keyword(s):  
Image Processing ◽  
Real Time ◽  
Target Detection ◽  
Ambient Noise ◽  
Arrival Direction ◽  
Processing Methods ◽  
Acoustic Target ◽  
Real Time Tracking ◽  
Passive Acoustic ◽  
Target Detecting

Passive acoustic target detection has been a hot research topic for a few decades. Azimuth recording diagram is one of the most promising techniques to estimate the arrival direction of the interested signal by visualizing the sound wave information. However, this method is challenged by the random ambient noise, resulting in low reliability and short effective distance. This paper presents a real-time postprocessing framework for passive acoustic target detection modalities by using a sonar array, in which image processing methods are used to automate the target detecting and tracking on the azimuth recording diagram. The simulation results demonstrate that the proposed approach can provide a higher reliability compared with the conventional ones, and is suitable for the constraints of real-time tracking.

scholarly journals A Research: Hyperspectral Image Processing Techniques

2019 ◽  
Vol 8 (9S2) ◽  
pp. 577-581
Keyword(s):  
Image Processing ◽  
Dimension Reduction ◽  
Target Detection ◽  
Hyperspectral Image ◽  
Target Identification ◽  
Target Classification ◽  
Detection Techniques ◽  
Hyperspectral Image Processing ◽  
Complicated Process ◽  
Processing Techniques

Hyperspectral image contains more information which are gathered from numerous narrow wavebands from one or more regions, and large amount of data are huddled. An basic problems in hyperspectral image processing are dimension reduction, target detection, target identification, and target classification. In this document, we reviewed the latest activities of target classification, most frequently used techniques for dimension reduction, target detection. Hyperspectral image processing is a complicated process which rely on mixed agents. Here we also recognized and reviewed problems faced by some methods and to overcome the problems, current techniques are discussed and highlighted good methods. To improving correctness, genuine classification techniques and Detection Techniques analysis are recommended

Design of a New Image Processing System Based on FPGA

2014 ◽  
Vol 1003 ◽  
pp. 216-220 ◽  
Author(s):  
Qi Li ◽  
Yu Yang ◽  
Zhong Ke Li ◽  
Jing Lu
Keyword(s):  
Image Processing ◽  
Unmanned Aerial Vehicles ◽  
Real Time ◽  
Target Detection ◽  
Processing System ◽  
Image Processing Algorithm ◽  
Verilog Hdl ◽  
Video Decoding ◽  
Image Processing System ◽  
Aerial Vehicles

According to the unmanned aerial vehicles real-time video image acquiring and target detection requirements, an image processing system was designed based on FPGA and TVP5150A decoder, and the video decoding hardware and software was also designed to meet the demands of unmanned aerial vehicles. An I2C controller was realized to assure the implementation of video decoding process in accordance with the requirements, and an image processing algorithm and applied to the image recognition process. Both of these were completed in FPGA using verilog HDL language. The correction of this image processing system was verified through real-time experiments.

scholarly journals CF2PN: A Cross-Scale Feature Fusion Pyramid Network Based Remote Sensing Target Detection

2021 ◽  
Vol 13 (5) ◽  
pp. 847
Author(s):  
Wei Huang ◽  
Guanyi Li ◽  
Qiqiang Chen ◽  
Ming Ju ◽  
Jiantao Qu
Keyword(s):  
Remote Sensing ◽  
Image Processing ◽  
Target Detection ◽  
Feature Fusion ◽  
Detection Methods ◽  
Natural Image ◽  
Scale Feature ◽  
Multi Scale ◽  
Backbone Network ◽  
Scale Detection

In the wake of developments in remote sensing, the application of target detection of remote sensing is of increasing interest. Unfortunately, unlike natural image processing, remote sensing image processing involves dealing with large variations in object size, which poses a great challenge to researchers. Although traditional multi-scale detection networks have been successful in solving problems with such large variations, they still have certain limitations: (1) The traditional multi-scale detection methods note the scale of features but ignore the correlation between feature levels. Each feature map is represented by a single layer of the backbone network, and the extracted features are not comprehensive enough. For example, the SSD network uses the features extracted from the backbone network at different scales directly for detection, resulting in the loss of a large amount of contextual information. (2) These methods combine with inherent backbone classification networks to perform detection tasks. RetinaNet is just a combination of the ResNet-101 classification network and FPN network to perform the detection tasks; however, there are differences in object classification and detection tasks. To address these issues, a cross-scale feature fusion pyramid network (CF2PN) is proposed. First and foremost, a cross-scale fusion module (CSFM) is introduced to extract sufficiently comprehensive semantic information from features for performing multi-scale fusion. Moreover, a feature pyramid for target detection utilizing thinning U-shaped modules (TUMs) performs the multi-level fusion of the features. Eventually, a focal loss in the prediction section is used to control the large number of negative samples generated during the feature fusion process. The new architecture of the network proposed in this paper is verified by DIOR and RSOD dataset. The experimental results show that the performance of this method is improved by 2–12% in the DIOR dataset and RSOD dataset compared with the current SOTA target detection methods.

Sign in / Sign up

Export Citation Format

Share Document