像素重要性测量特征下的侧扫声呐目标检测
|
摘要
研究了在较低信噪比下,在保证检测概率的前提下尽量降低虚警概率的目标检测,提出了一种针对特定目标的两阶段筛选算法.第一阶段中,首先使用阈值分割出有效点,并定义了一种新的像素重要性测量特征用于初步筛选目标。即通过有效像素点之间的距离来赋以高斯分布的权值,当前像素重要性的值定义为剩余有效点的距离加权和,具有较高的像素重要性值的聚集性强的区域内像素点会被定位出来。第二阶段,使用卷积神经网络分类器排除虚假目标.在实验中,使用近期无人潜器获得的海底数据,召回率与虚警概率分别达到90.39%与2.39%,证明了其相比声呐目标检测主流算法有更好的检测能力。
The prior task is to stabilize the detection ability and reduce false alarm probability simultaneously underlow signal-noise ratio condition. A two-stage novel pixel importance measurement algorithm in the side-scan sonar detection application is proposed. In first stage of the algorithm, a new feature defined as Pixel Importance Value(PIV) is proposed based on distances between the target pixel and each other pixels. PIV measurement of current pixel is defined as the weighted sum of all remaining segmented pixels. The weighted part refers to Gaussian kernel, which means closer pixels gets higher weight. Thus, targets with higher pixel importance value can be located. In the second stage, we use convolutional neural network as classifier to eliminate the dot-like false targets in terms of targets shape. In our experiments based on recent underwater data obtained by autonomous underwater vehicle, we demonstrate superior performance of our algorithm over the state-of-the-art sonar detection algorithms in terms of 90.39% recall rate and 2.39% false alarm probability.
The prior task is to stabilize the detection ability and reduce false alarm probability simultaneously underlow signal-noise ratio condition. A two-stage novel pixel importance measurement algorithm in the side-scan sonar detection application is proposed. In first stage of the algorithm, a new feature defined as Pixel Importance Value(PIV) is proposed based on distances between the target pixel and each other pixels. PIV measurement of current pixel is defined as the weighted sum of all remaining segmented pixels. The weighted part refers to Gaussian kernel, which means closer pixels gets higher weight. Thus, targets with higher pixel importance value can be located. In the second stage, we use convolutional neural network as classifier to eliminate the dot-like false targets in terms of targets shape. In our experiments based on recent underwater data obtained by autonomous underwater vehicle, we demonstrate superior performance of our algorithm over the state-of-the-art sonar detection algorithms in terms of 90.39% recall rate and 2.39% false alarm probability.
引文
1 Rahmati M, Pompili D. UNISeC:Inspection, separation, and classification of underwater acoustic noise point sources. IEEE J. Oceanic Eng., 2018; 43(3):777-791
2王晓,吴清海,王爱学.侧扫声呐图像辐射畸变综合改正方法研究.大地测量与地球动力学,2018; 38(11):1174-1179
3徐东,李风华,郭永刚,王元.海面波浪谱对深海低频环境噪声的影响.声学学报,2018; 43(2):137-144
4 Bian H Y, Chen Y M, Liu W J. Circular target detection in noisy sonar image. In:Proceedings of 12th Western Pacific Acoustics Conference, Singapore, 2015:1-8
5 Maussang F, Chanussot J, Hetet A et al.Higher-order statistics for the detection of small objects in a noisy background application on sonar imaging. EURASIP J. Adv.Signal Process., 2007; 2007(1):1-17
6王丽娜.侧扫声呐目标特征提取方法研究.哈尔滨工程大学,2014
7 Coroyer C, Jorand C, Duvaut P. ROC curves of skewness and kurtosis statistical tests:application to textures. In:Proceedings of the 7th European Signal Processing Conference(EUSIPCO'94), Edinburgh, Scotland, UK, 1994:450-453
8 Aviles-Cruz C, Rangel-Kuoppa R, Reyes-Ayala M et al.High-order statistical texture analysis-font recognition applied. Pattern Recogn. Lett.,2005; 26(2):135-145
9 Rajagopalan A N, Jain A, Desai U B. Data clustering using hierarchical deterministic annealing and higher order statistics:Image processing. IEEE Transactions on CirCcuits&Systems II Analog&Digital Signal Processing,1999; 46(8):1100-1104
10张石磊,胡志成,付丽华.基于高阶统计量的相干算法研究.地球物理学进展,2016; 31(4):1762-1766
11 Carrato S, Ramponi G. Edge detection using generalized higher-order statistics. IEEE Signal Processing Workshop on Higher-order Statistics, IEEE, 1993:66-70
12杨信能,朱宝山,李润生,李雪英.独立分量分析及其在图像处理方面应用.影像技术,2013; 25(5):48-50
13 Ramponi G,Carrato S,Performance of the Skewness-ofGaussian(SoG)edge extractor. In:Proceedings of the 7th European Signal Processing Conference(EUSIPCO'94),Edinburgh, Scotland, UK, 1994:454-457
14 Li H, Gao J, Du W et al. Object representation for multi-beam sonar image using local higher-order statistics.EURASIP J. Adv. Signal Process., 2017; 2017(1):7
15 Kaeli J W. Real-time anomaly detection in side-scan sonar imagery for adaptive AUV missions. Autonomous Underwater Vehicles. IEEE, 2016:85-89
16 Williams D P. The mondrian detection algorithm for sonar imagery. IEEE Trans. Geosci. Remote Sens., 2018(56):1091-1102
17 Nicolas F, Arnold-Bos A, Quidu I et al. Markov-based approaches for ternary change detection between two highresolution synthetic aperture sonar tracks. Oceans, Aberdeen, Scotland, 2017
18谢小敏.水下图像分割和典型目标特征提取及识别技术研究.南京理工大学,2015
19 Liu X, Xiong Z. Underwater small target tracking algorithm based on diver detection sonar image sequences. International Conference on Industrial Control and Electronics Engineering. IEEE, Saint Joseph, USA, 2012:727-730
20 Bardhan S, Rajapan D, Zacharia S et al. Detection of buried objects using active sonar. Underwater Technology,IEEE, 2015:1-5
21 Wang X, Liu S, Teng X et al. SFLA with PSO local search for detection sonar image. Control Conference, IEEE,Boston, MA, USA, 2016:3852-3857
22王雷,叶秀芬,王天.模糊聚类的侧扫声呐图像分割算法.华中科技大学学报(自然科学版),2012; 40(9):25-29
23罗明愿.侧扫声呐图像感兴趣区域提取算法研究.哈尔滨工程大学,2011
24罗明愿,卞红雨,周志娟.基于图像频域方向模板的目标检测方法.计算机工程,2011; 37(12):215-217
25 Wang X, Liu S, Sun J et al. A novel quantum genetic algorithm for detection sonar image. Control and Decision Conference. IEEE, Boston, MA, USA, 2016:1991-1996
26 Kim J, Cho H, Pyo J et al.The convolution neural network based agent vehicle detection using forward-looking sonar image. Oceans, IEEE, Monterey, CA, USA, 2016:1-5
27朱兆彤,彭石宝,许稼,许肖梅.水下目标亮点拓扑特征提取及自动识别方法.声学学报,2018; 43(2):154-162
28张珊.基于PCA-SURF的单目视觉SLAM算法.中国海洋大学,2014
29 Ioffe S, Szegedy C. Batch normalization:Accelerating deep network training by reducing internal covariate shift. International Conference on International Conference on Machine Learning. JMLR.org, Lille, France, 2015:448-456
2王晓,吴清海,王爱学.侧扫声呐图像辐射畸变综合改正方法研究.大地测量与地球动力学,2018; 38(11):1174-1179
3徐东,李风华,郭永刚,王元.海面波浪谱对深海低频环境噪声的影响.声学学报,2018; 43(2):137-144
4 Bian H Y, Chen Y M, Liu W J. Circular target detection in noisy sonar image. In:Proceedings of 12th Western Pacific Acoustics Conference, Singapore, 2015:1-8
5 Maussang F, Chanussot J, Hetet A et al.Higher-order statistics for the detection of small objects in a noisy background application on sonar imaging. EURASIP J. Adv.Signal Process., 2007; 2007(1):1-17
6王丽娜.侧扫声呐目标特征提取方法研究.哈尔滨工程大学,2014
7 Coroyer C, Jorand C, Duvaut P. ROC curves of skewness and kurtosis statistical tests:application to textures. In:Proceedings of the 7th European Signal Processing Conference(EUSIPCO'94), Edinburgh, Scotland, UK, 1994:450-453
8 Aviles-Cruz C, Rangel-Kuoppa R, Reyes-Ayala M et al.High-order statistical texture analysis-font recognition applied. Pattern Recogn. Lett.,2005; 26(2):135-145
9 Rajagopalan A N, Jain A, Desai U B. Data clustering using hierarchical deterministic annealing and higher order statistics:Image processing. IEEE Transactions on CirCcuits&Systems II Analog&Digital Signal Processing,1999; 46(8):1100-1104
10张石磊,胡志成,付丽华.基于高阶统计量的相干算法研究.地球物理学进展,2016; 31(4):1762-1766
11 Carrato S, Ramponi G. Edge detection using generalized higher-order statistics. IEEE Signal Processing Workshop on Higher-order Statistics, IEEE, 1993:66-70
12杨信能,朱宝山,李润生,李雪英.独立分量分析及其在图像处理方面应用.影像技术,2013; 25(5):48-50
13 Ramponi G,Carrato S,Performance of the Skewness-ofGaussian(SoG)edge extractor. In:Proceedings of the 7th European Signal Processing Conference(EUSIPCO'94),Edinburgh, Scotland, UK, 1994:454-457
14 Li H, Gao J, Du W et al. Object representation for multi-beam sonar image using local higher-order statistics.EURASIP J. Adv. Signal Process., 2017; 2017(1):7
15 Kaeli J W. Real-time anomaly detection in side-scan sonar imagery for adaptive AUV missions. Autonomous Underwater Vehicles. IEEE, 2016:85-89
16 Williams D P. The mondrian detection algorithm for sonar imagery. IEEE Trans. Geosci. Remote Sens., 2018(56):1091-1102
17 Nicolas F, Arnold-Bos A, Quidu I et al. Markov-based approaches for ternary change detection between two highresolution synthetic aperture sonar tracks. Oceans, Aberdeen, Scotland, 2017
18谢小敏.水下图像分割和典型目标特征提取及识别技术研究.南京理工大学,2015
19 Liu X, Xiong Z. Underwater small target tracking algorithm based on diver detection sonar image sequences. International Conference on Industrial Control and Electronics Engineering. IEEE, Saint Joseph, USA, 2012:727-730
20 Bardhan S, Rajapan D, Zacharia S et al. Detection of buried objects using active sonar. Underwater Technology,IEEE, 2015:1-5
21 Wang X, Liu S, Teng X et al. SFLA with PSO local search for detection sonar image. Control Conference, IEEE,Boston, MA, USA, 2016:3852-3857
22王雷,叶秀芬,王天.模糊聚类的侧扫声呐图像分割算法.华中科技大学学报(自然科学版),2012; 40(9):25-29
23罗明愿.侧扫声呐图像感兴趣区域提取算法研究.哈尔滨工程大学,2011
24罗明愿,卞红雨,周志娟.基于图像频域方向模板的目标检测方法.计算机工程,2011; 37(12):215-217
25 Wang X, Liu S, Sun J et al. A novel quantum genetic algorithm for detection sonar image. Control and Decision Conference. IEEE, Boston, MA, USA, 2016:1991-1996
26 Kim J, Cho H, Pyo J et al.The convolution neural network based agent vehicle detection using forward-looking sonar image. Oceans, IEEE, Monterey, CA, USA, 2016:1-5
27朱兆彤,彭石宝,许稼,许肖梅.水下目标亮点拓扑特征提取及自动识别方法.声学学报,2018; 43(2):154-162
28张珊.基于PCA-SURF的单目视觉SLAM算法.中国海洋大学,2014
29 Ioffe S, Szegedy C. Batch normalization:Accelerating deep network training by reducing internal covariate shift. International Conference on International Conference on Machine Learning. JMLR.org, Lille, France, 2015:448-456