Fengqi Zhang

Fengqi Zhang

Auto-generated profile, based on publicly available information

Fengqi Zhang

State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, China

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
2 July 2024 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
FZFengqi Zhang
Fengqi Zhang
ZTZhigang Tu
Zhigang Tu
WHWeifeng Hao
Weifeng Hao
YCYihao Chen
Yihao Chen
FLFei Li
Fei Li
Mao YeMao Ye
Mao Ye
Obtaining high-visibility images of the lunar polar permanently shadowed region (PSR) is quite important for internal landforms and material existence exploration. However, PSR images are usually have poor quality due to lack of sufficient illumination. Existing researches, which attempt to address this problem, face challenges caused by relying on virtual assumptions, manual processing, and paired data. To solve these problems, we aim to avoid using paired datasets and directly optimize PSR image, and accordingly propose a zero-shot parameter learning model (ZSPL-PSR) for PSR image enhancement. Our ZSPL-PSR, which enhances PSR images by estimating parameters to adjust image properties, consists of a parameter learning network and a parameter weight learning structure. Particularly, first, a parameter learning network that integrates robust information is constructed to separately estimate the mid-tone brightness parameters, shadow brightness parameters, and contrast parameters. Where these parameters are beneficial for iteratively improve the overall brightness, shadow brightness, and contrast of the image. Second, a parameter weight learning structure is exploited to coordinate the priority of different parameter maps. In addition, to highlight the terrain details in the enhanced PSR image, we use USM sharpening for post-processing. The experimental results display the fully inter-pretable enhanced PSR maps of the lunar north and south poles and their sharpened versions, showcasing rich landforms in PSR. To validate the model performance, a benchmark PSR testing set has constructed, and extensive comparisons conducted on it demonstrated that ZSPL-PSR exceeds other zero-shot learning methods significantly in image quality. Our code is available at https://github.com/dl-zfq/ZSPL-PSR.

28 February 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
FZFengqi Zhang
Fengqi Zhang
WHWeifeng Hao
Weifeng Hao
Mao YeMao Ye
Mao Ye
ZTZhigang Tu
Zhigang Tu
FLFei Li
Fei Li
Cameras carried by satellites take large-scale images of the planetary surface, which provides data support for exploration missions. With the progress of deep space exploration, the demands to identify surface changes are gradually increasing. There are few studies focus on planetary surface automatic change detection, since it faces the challenges of difficult registration, strong pseudo change, poor automatic detection capacity, etc. To handle these challenges, we propose a deep learning lightweight model SLCD-Net (i.e., striped landform change detection model) for patch-wise change detection of the planetary striped landform. SLCD-Net uses the dual-input-based siamese network as the overall architecture to learn the deep semantic information of pre- and post-temporal images. Importantly, SLCDNet designs a dual-feature multi-level complementary fusion module between two branches of the siamese network, which can learn the cross-temporal complementary features and eliminate the noise impact. Besides, SLCD-Net exploits a spatial attention module after the fusion unit, which further helps to weaken noise that is produced by fusion. Furthermore, we construct a dataset about dark slope streak (DSS) (a striped landform type) ( https://github.com/dl-zfq/Mars-CD-dataset ). Training on our DSS dataset enables SLCD-Net to have an advantage in identifying the variation of DSS. The SLCD-Net's detection performance on multiple planetary surface datasets including our DSS test set proves that it has achieved a breakthrough in robustness and generalization. Not limited to the patch-wise task, the change map generated by SLCD-Net applied to different regions also proves that it can be effectively used for pixel-wise Mars striped landform change detection.

23 April 2022 in International Journal of Pattern Recognition and Artificial Intelligence
Verified
Journal Article
Closed access
PDF file
LYLina Yang
Lina Yang
FZFengqi Zhang
Fengqi Zhang
PWPatrick Shen-Pei Wang
Patrick Shen-Pei Wang
XLXichun Li
Xichun Li
HLHuiwu Luo
Huiwu Luo

9 September 2025 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
FZFengqi Zhang
Fengqi Zhang
Mao YeMao Ye
Mao Ye
WHWeifeng Hao
Weifeng Hao
XSXuemei Sun
Xuemei Sun
FLFei Li
Fei Li
The craters widely distributed on the lunar surface constitute important landforms for investigating the geological activities and evolutionary history of the Moon. Currently, deep learning models have made certain progress in detecting craters, but they still have limitations in identifying numerous small craters. To address these challenges, this paper puts forward a CenterNet-based dense small crater detection model (DSCraterNet), which is designed to precisely detect craters within the permanently shadowed regions (PSRs) by leveraging the previously released enhanced PSR images. DSCraterNet achieves precise crater localization by predicting keypoint heatmap (location), size, and offset of craters. The model employs an encoder-decoder network architecture that combines various feature fusion and attention mechanisms. Feature fusion strategies include direct fusion between the encoder and decoder, multi-scale encoder spatial feature fusion structure (MSF), and multi-element information fusion structure (MEF). These fusion strategies balance the relationship between high-resolution features, strong semantic features, and multi-scale spatial features according to craters’ characteristics. Furthermore, by reasonably configuring attention-based layers (like efficient multi-scale attention (EMA)) at different stages, the utilization efficiency of robust features is improved. Ultimately, DSCraterNet successfully identified a large number of densely packed small craters in PSR, revealing the terrain features of the region. A series of thorough ablation and comparative experiments have confirmed that DSCraterNet surpasses existing crater detection models in terms of precision, recall, mAP, and F1-score metrics. Our code is available at https://github.com/dl-zfq/DSCraterNet.

1 January 2026 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
QYQiu Yi
Qiu Yi
YQYue Que
Yue Que
Jie YouJie You
Jie You
FZFengqi Zhang
Fengqi Zhang
NJNan Jiang
Nan Jiang
YYYong Yang
Yong Yang

3 July 2024
Verified
Conference Paper
Open access
FZFengqi Zhang
Fengqi Zhang
MYMao Ye
Mao Ye
WHWeifeng Hao
Weifeng Hao
YCYihao Chen
Yihao Chen
FLFei Li
Fei Li
The moon is the earth's only natural satellite and it has a huge impact on the earth. Many countries have carried out a large number of lunar exploration missions over the years. The Lunar Reconnaissance Orbiter (LRO) launched by the United States in 2009 carried seven scientific instrument payloads to comprehensively map the lunar terrain and explore lunar resources. The optical narrow-angle camera (NAC) captures, processes and transmits back a large amount of image data [1]. These optical image data are intuitive data for studying the moon. Due to the axial tilt and undulating terrain, there are many areas (i.e., permanently shadowed regions (PSRs)) near the lunar south pole and north pole that are never illuminated by sunlight, so the imaging quality of these areas is poor. As shown in Fig. 1, PSR images have low signal-to-noise ratio (SNR), poor visibility, and most of the content is obscured. However, PSR may contain important materials such as water-ice, which makes it critical to study PSR with images. How to use low-light image enhancement methods to obtain high-quality PSR images has become a major challenge in this field.Fig. 1. Optical image examples of the lunar south pole PSR. These images were captured by LRO NAC and stitched together in post-processing.Image enhancement is an important task in computer vision, which aims to design algorithms to improve the interpretability of images. Image enhancement in the past was mainly divided into traditional methods and deep learning methods. Traditional methods represented by histogram normalization (HE)[2] and dehazing model theory[3] are limited in enhancement range and physical assumptions, and have low automation capabilities. Deep learning methods are divided into supervised learning methods, GAN-based methods[4] and zero-shot learning methods. Supervised learning relies on pre-made paired datasets consisting of original and optimized images. However, for PSR images, there is no standard optimized image, so supervised learning methods are not suitable for PSR image enhancement. Although GAN-based methods get rid of the restrictions on paired datasets, they still need to select data to meet the needs of the generator and discriminator, which makes the process more cumbersome. The zero-shot learning method, which does not rely on any data and can directly enhance images, is suitable for the enhancement of scarce data such as PSR images. An important category of zero-shot learning methods is the decomposition of low-quality images derived from Retinex decomposition theory. Regarding an image G(x), Retinex theory believes that it can be divided into a combination of reflection map and illumination map, that is:G(x) = R(x)•I(x)where x is the pixel position of the image, R and I represent the reflection map and illumination map respectively. For low-light images such as PSR images, I(x) is too dark, and R, which represents image features, should also be adjusted. By adjusting these sub-maps, the Retinex-based method can achieve image enhancement.However, traditional Retinex only considers the content and brightness of the image (i.e., reflectance map and illumination map), but does not consider image noise. Low-light images such as PSR have a low number of photons transmitted to the CCD during imaging, resulting in dim visual phenomena. But low photons not only reduce the brightness of the image, but also introduce a lot of noise, which further reduces the quality of the image. The traditional Retinex method will expand the original noise while adjusting the brightness, which hinders image content analysis. To this end, we propose a Retinex-based model RZSL-PSR that introduces noise estimation. This model decomposes the PSR image into reflection map, illumination map and noise map. The specific formula is:G(x) = R(x)•I(x) + N(x)Based on this formula, the model uses the proposed sub-map learning network to output three sub-maps respectively. The network structure includes downsampling, concatenation fusion, and skip connection structures, which helps to obtain feature maps that integrate multi-scale, strong semantic information, and high-resolution information. Subsequently, the three output sub-maps are optimized to obtain the final output enhanced image. This process does not rely on learning image mappings of paired datasets, but instead utilizes a non-reference loss function to supervise model training. The loss function is as follows:L = Lr + a•Lt + b•Lnwhere Lr is the loss function to ensure that the output image conforms to the Retinex decomposition theory, Lt is the loss function to flatten the illumination uniformity of the image, and Ln is the loss function to estimate and suppress the noise in the dark area. Guided by these loss functions, the model learns nonlinear function mappings for enhancing PSR images, including increasing brightness and removing noise.As shown in Fig. 2, the enhanced PSR image has better visibility than the original image. However, in order to highlight the landforms within the area for research on landforms and materials, we use USM sharpening for processing. The sharpening map results in Fig. 2 show that the landforms inside the PSR are highlighted, which helps to count the internal landforms.Fig. 2. PSR’s (original image)-(enhanced image)-(sharpened image) comparison example. Among them (a) is the input PSR image, (b) is the enhancedimage, and (c) is the image after USM sharpening of the enhanced image.AcknowledgementsThis work was financially supported by the National Natural Science Foundation of China under No. 42030110, 41874010 and 41804025, the Natural Science Foundation of Hubei Province of China under No. 2022CFB123 and the Space Optoelectronic Measurement and Perception Lab in Beijing Institute of Control Engineering under No. LabSOMP-2023-09.References[1] Vondrak R, Keller J, Chin G, et al. Lunar Reconnaissance Orbiter (LRO): Observations for lunar...