Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization
Open accessThe DOI matches its Crossref record; 48 of the 48 references it cites were found in a registry or an index; 2 of the 9 sources its claims cite were checked for support.
Abstract
Radiation anomalies in optical remote sensing images frequently occur due to electronic issues within the image sensor or data transmission errors. ... To ensure the retention of image data with minimal radiation issues as much as possible, this paper adopts a self-made radiation dataset and proposes a FlexVisionNet-YOLO network to detect radiation anomalies more accurately. Firstly, RepViT is used as the backbone network with a vision transformer architecture to better capture global and local features. ...
Overview
CoChat Summary
TL;DR
The paper proposes FlexVisionNet-YOLO, a YOLOv8-based detector with a RepViT backbone, dynamic deep feature fusion (SDI, Dy_Sample), Inner-CIoU, and Slide-Loss for radiation anomaly detection in sub-band optical remote sensing images. On a self-made six-class radiation anomaly dataset, it reports higher precision, recall, and mAP than YOLOv8 and other compared detectors.
Key findings
- FlexVisionNet-YOLO achieved P 0.982, R 0.952, mAP@0.5 0.973, and mAP@0.5:0.9 0.901, compared with YOLOv8’s P 0.947, R 0.881, mAP@0.5 0.929, and mAP@0.5:0.9 0.765; the paper states improvements of 3.5%, 7.1%, 4.4%, and 13.6%, respectively.
- Category-wise, FlexVisionNet-YOLO reported higher values than YOLOv8, e.g., Image-Missing P 0.993 vs. 0.934, RandomCode1 P 0.965 vs. 0.914, CCD R 0.967 vs. 0.842, and RandomCode1 mAP@0.5:0.9 0.834 vs. 0.633.
- Ablation results show RepViT reduced inference time from 5.8 ms to 2.3 ms in Method 1 while improving recall; adding Inner-CIoU and Slide-Loss improved recall by 5.3%; the final model ran at 3.7 ms.
- On 100 real images from ZY-3, GF1, GF2, GF6, and GF7, the model maintained a false detection rate of about 5%; band splitting produced clearer detection results and improved accuracy.
Methods
- A self-made radiation anomaly dataset was built from ZY3-02, GF-1, GF-2, GF-6, and GF-7 data: more than 800 multispectral and panchromatic images were cropped to 640 × 640, and 8300 images were used for training. The six classes are CCD, StripeNoise, RandomCode1, RandomCode2, ImageMissing, and Tap; RGB bands were split into independent images for analysis.
- The model modifies YOLOv8 by using RepViT as the backbone, replacing the Neck’s Concat with SDI and up-sampling with Dy_Sample in a Dynamic Deep Feature Fusion Network, and using Inner-CIoU for box regression and Slide-Loss for classification loss.
- Experiments used an Intel Xeon E5-2650 v4 CPU, NVIDIA Titan V GPU, Windows 10 Professional, and PyTorch 1.8; evaluation metrics were Precision, Recall, mAP@0.5, mAP@0.5:0.9, and GFLOPs, with comparisons against SSD, Faster R-CNN, YOLOv5, YOLOv6, YOLOv7, and YOLOv8.
Limitations
- Stated: the samples were preprocessed to eliminate some bias color problems, but bias color problems still exist in practical applications and need further solving.
- Stated: further optimization of the model’s computational efficiency remains a key direction for future research.
- Evident from the text: evaluation relies on a self-made dataset and only 100 real images for testing, and the data are available from the authors rather than a public benchmark, which may limit reproducibility and generalization assessment.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures













Verification
Claims the paper makes
2 Valid · 8 Not valid · 7 Not checked · 97 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“As shown in Figure 7, Inner-CIoU [45] can more accurately assess the overlap region between the predicted and real boxes by introducing centroid consistency and shape consistency, which improves the accuracy of bounding box localization.”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 2.6. Inner-CIoU Loss and Slide Loss Function
What the sources say
- Partly supportedcites [45] Zhang, H., Xu, C., and Zhang, S. (2023). Inner-iou: More effective intersection over union loss with auxiliary bounding box. arXiv. sourceExperimental — accuracy not yet measured
The Introduction (page not provided) describes CIoU as combining a center-distance term with a shape-consistency term, and the YOLOv7 on PASCAL VOC section reports more accurate positioning with Inner-CIoU. But the Inner-IOU Loss section says Inner-CIoU adds an auxiliary-box IoU term to CIoU; it does not introduce the center and shape terms as the reason for its improvement. (LLM, conf=0.86)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“This method overcomes the limitations of existing approaches by using auxiliary bounding boxes of different scales for the dataset and detector, thereby improving its generalization capabilities.”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 2.6. Inner-CIoU Loss and Slide Loss Function
What the sources say
- Supportedcites [45] Zhang, H., Xu, C., and Zhang, S. (2023). Inner-iou: More effective intersection over union loss with auxiliary bounding box. arXiv. source
The “B. Inner-IOU Loss” section explicitly says that using different-scale auxiliary bounding boxes for different datasets and detectors can overcome the weak generalization of existing methods. The conclusion also reports that experiments confirmed the method’s generalization. The supplied passages have section labels but no page numbers. (LLM, conf=0.94)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The experimental results demonstrate that the network still maintains high performance in real application scenarios.”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 1. Introduction
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Its multiscale feature fusion mechanism improves target detection of different sizes and shapes”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 1. Introduction
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the optimized self-attention mechanism significantly improves adaptability in handling complex scenes”
- Also stated in: § 5. Conclusions · plain
“The multiscale feature extraction mechanism enhances the detection of targets of different sizes and shapes, while the optimized self-attention mechanism significantly improves adaptability in handling complex scenes.”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 1. Introduction
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § 5. Conclusions · plain
42 more claims
Create a free account to see all 114 claims and every source verdict.
- Probably not a substantial claim. 23
- It may state two separate claims as one. 55
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 3.2. Analysis of Experimental Results of Different Models
“YOLOv8, on the other hand, performs better in precision (0.947), recall (0.881), and mAP@0.5 (0.765).”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 5. Conclusions
“| Yolov8 | 0.947 | 0.881 | 0.929 | 0.765 |”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 3.1. Experimental Environment Setting and Evaluation Metrics
“AP in Equation ( 7) denotes the average precision of category i.The variable "n" in Equation ( 8) represents the overall amount of categories, and averaging the APs of n categories gives mAP.”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 3.1. Experimental Environment Setting and Evaluation Metrics
“P = TP TP + FP (6 ) R = TP TP + FN (7 ) AP i = 1 0 P(R)dR (8 ) mAP = 1 n ∑ n i=1 AP i (9”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 3.1. Experimental Environment Setting and Evaluation Metrics
“AP in Equation ( 7) denotes the average precision of category i.The variable "n" in Equation ( 8) represents the overall amount of categories”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 3.1. Experimental Environment Setting and Evaluation Metrics
“R = TP TP + FN (7 ) AP i = 1 0 P(R)dR (8 ) mAP = 1 n ∑ n i=1 AP i (9”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 1. Introduction
“Etten et al. [29] introduced the YOLOv2 model and proposed the YOLT method”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > References
“29. Law, H.; Deng, J. CornerNet: Detecting Objects as Paired Keypoints. Int. J. Comput. Vis. 2020, 128, 642-656. [CrossRef]”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 1. Introduction
“SDI (Spatial Dynamic Integration)”
§ Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization > 2.5. Dynamic Deep Feature Fusion Networks
“SDI (Selective Deconvolution Integration)”
Citations
Cites
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1506.01497 · Cited by 70,812 · cited 1×
“The two-stage detection model was first proposed by Girshick et al. [8], which detects targets by searching for regions in the image that may contain objects, with representative algorithms such as Fast R-CNN [9], SPP-net [10], and Faster R-CNN [11].”
MethodYou Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · cited 1×
“To solve this problem, single-stage target detectors have emerged [12], which directly process the entire image and complete the detection task with a single network.”
BackgroundGirshick, R. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.169 · Cited by 28,105
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2389824 · Cited by 11,971
He, K., Gkioxari, G., Dollar, P., & Girshick, R. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2018.2844175 · Cited by 3,457
DOTA: A Large-Scale Dataset for Object Detection in Aerial Images
Xia, G.-S., Bai, X., Ding, J., Zhu, Z., Belongie, S., Luo, J., Datcu, M., Pelillo, M., & Zhang, L. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00418 · Cited by 3,414
Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10578-9_23 · Cited by 3,034
Region-based convolutional networks for accurate object detection and segmentation
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2016) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2437384 · Cited by 2,912
Deep learning in environmental remote sensing: Achievements and challenges
Yuan, Q., Shen, H., Li, T., Li, Z., Li, S., Jiang, Y., Xu, H., Tan, W., Yang, Q., Wang, J., et al. (2020) · Remote Sensing of Environment · doi:10.1016/j.rse.2020.111716 · Cited by 1,697
Cited by
Research on object detection and recognition in remote sensing images based on YOLOv11
He, L.-H., Zhou, Y.-Z., Liu, L., Cao, W., & Ma, J.-H. (2025) · Scientific Reports · doi:10.1038/s41598-025-96314-x · Cited by 208
Chen, M., Zhang, H., Cheng, B., Li, K., Li, Y., Guo, C., Liu, J., Li, Y., Jing, L., & Fang, X. (2025) · International Journal of Remote Sensing · doi:10.1080/01431161.2025.2564908 · Cited by 1
“Shen et al. (Ci et al. 2024) proposed a novel dynamic sensing and correlation loss detector (DCDet) for multi-scale object detection.”
(2026) · doi:10.1109/tim.2026.3687318
“Ci et al. [16] introduce a radiation anomaly detection method based on multiscale deep dynamic fusion and adaptive optimization for optical remote sensing images.”
Broadband polarization-insensitive achromatic metalens with high focusing efficiency
Weng, Q., Ouyang, X., Li, Y., Fu, S., Ding, Z., Hu, J., Shen, J., & Li, C. (2025) · Optik · doi:10.1016/j.ijleo.2025.172558 · Cited by 1
Metadata
- Authors
- Ci, J., Tan, H., Zhai, H., & Tang, X.
- Affiliations
- Ministry of Natural Resources (CN)
- Corresponding author
- Tan, Hai
- Venue
- Remote Sensing
- Publisher
- MDPI AG
- Published
- 2024-08-12
- Volume
- 16
- Issue
- 16
- Pages
- 2953
- ISSN
- 2072-4292 (ISSN-L)
- OpenAlex
- W4401510301
- Semantic Scholar
- 64937096f994c144948d29cc5ad03380bae4d9ad
- Cited by
- 4
- References
- 43
- Citations by year
- 2026: 1 · 2025: 3
- FWCI
- 0.66
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Remote Sensing (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- Demonstration System for High-Resolution Remote Sensing and Mapping Applications (Phase II) (42-Y30B04-9001-19/21, 00000113)
- Construction of natural resources satellite remote sensing technology system and application demonstration (42-Y30B04-9001-19/21, 00000113)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Remote-Sensing Image Classification
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Industrial Vision Systems and Defect Detection
- Keywords
- multi-scale feature fusion, RepViT, vision transformer, YOLO, stripe noise, sample imbalance, object detection and classification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (prompt extract-v3).
A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization. https://app.cochat.ai/papers/10.3390/rs16162953
- The paper
- Ci, J., Tan, H., Zhai, H., & Tang, X. (2024). Radiation Anomaly Detection of Sub-Band Optical Remote Sensing Images Based on Multiscale Deep Dynamic Fusion and Adaptive Optimization. Remote Sensing. https://doi.org/10.3390/rs16162953
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.