Application of spectral imaging technology in maritime target detection

Feng-hua, 梅风华 Mei; Chao, 李超 Li; Yu-xin, 张玉鑫 Zhang

doi:10.3788/co.20171006.0708

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The input image size for all methods is M × N × L. In our method, b denotes the number of nodes in the latent space, while iter represents the number of layers in hCEM, which is on average 10. For CSCR, (w o ut, w i n) form the sliding double windows and are set to (11,3). The values of ncem and nlayer represent the number of CEM detectors used per layer and the number of layers in ECEM, respectively, with values of 6 and 10.…”

Section: Compared Methodsmentioning

confidence: 99%

“…The spectral information therefore reflects the differences in the internal structure and composition of target samples. These distinct features make hyperspectral imaging technology applicable in various fields, including mineral detection [1], disease diagnosis [2], military reconnaissance [3], and environmental detection [4]. Long Sun, Zongfang Ma, and Yi Zhang are with the College of Information and Control Engineering, Xi'an University of Architecture and Technology, Xi'an 710311, China (e-mail: sunlongxauat@outlook.com; mazf@xauat.edu.cn; zhangyi@xauat.edu.cn).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

ABLAL: Adaptive Background Latent Space Adversarial Learning Algorithm for Hyperspectral Target Detection

Sun,

Ma,

Zhang

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Hyperspectral images (HSIs) are challenging for hyperspectral object detection (HTD) due to their complex background information and the limited prior knowledge of the target. This paper proposes an adaptive background latent space adversarial learning algorithm for hyperspectral target detection (ABLAL). We begin by using a coarse screening method to select pseudo-background and pseudo-target sample sets, addressing the issues caused by insufficient prior target information and complicated background information, which result in low detection accuracy. Next, we utilize an Adversarial Autoencoder (AAE) based backbone network to extract the background latent spatial information of the HSI. It should be noted that we adaptively constrain the accuracy of the extracted information through the pseudo-target dataset, accounting for the impact of potential targets in the pseudo-background dataset. Furthermore, we fully utilize the information extracted by AAE and employ a strategy combining multiple output results of AAE. Specifically, we use the distance between the target latent space vector and the background latent space vector, and the HSI reconstruction difference to suppress the background. Finally, extensive experiments are conducted on real datasets to demonstrate the effectiveness of the proposed method.

show abstract

Section: Compared Methodsmentioning

confidence: 99%