Composite Georeferencing Method for Image Processing

Abstract

Image georeferencing is the process of accurately mapping an image onto a geographic coordinate system. It is really crucial for applications in remote sensing, urban planning, environmental analysis, and historical research. Numerous studies have evaluated georeferencing methods for their accuracy, efficiency, and application suitability. This paper introduces a new approach, the Composite Georeferencing Method (CGM). The CGM begins by applying an Affine or Helmert transformation to align the image control points with their real-world geographic coordinates. Next, a stepwise interpolation process is implemented, utilizing three proposed influence factors: Linear, Cosine, and Tangent. Results indicate that CGM, using these influence factors, yields better outcomes compared to traditional Helmert and Affine methods. Within the first quarter of the maximum distance between points, the Cosine and Tangent Influence Factors show similar performance, while the Linear Influence Factor proves to be less effective. Beyond this point, the Cosine Influence Factor experiences a rapid decline in its effect, which enhances overall performance. In the final quarter of the distance range, the Cosine Influence Factor flattens, becoming negligible, indicating minimal influence of point displacement at longer distances.

Introduction

1. Importance of Image Georeferencing

• Georeferencing assigns real-world geographic coordinates (like latitude and longitude) to each pixel in an image.

• It is a foundational tool for remote sensing, urban planning, environmental analysis, and archaeological studies.

• Enables accurate spatial analysis, integration with other datasets, and supports informed decision-making in global challenges.

• Georeferenced images improve spatial accuracy, allow overlays with vector data, and are critical for land-use monitoring, urban sprawl analysis, and infrastructure planning.

2. Applications in Various Fields

• Remote Sensing: Aligns satellite imagery with other datasets to monitor environmental change.

• Urban Planning: Supports tracking urban growth and managing land use.

• Natural Resource Management: Enables monitoring deforestation, water resources, etc.

3. Georeferencing Techniques

• Manual Georeferencing: User selects control points manually. Essential for high-precision applications like historical maps, though time-consuming.

• Automatic Georeferencing: Uses image processing and feature detection algorithms (e.g., SIFT, SURF) to align images automatically.

• Machine Learning Techniques:

  • Use of CNNs (Convolutional Neural Networks) enhances accuracy and automates georeferencing.

  • Offers robustness in complex environments, reducing manual work and increasing efficiency.

4. Common Software Tools

• ArcGIS, QGIS, ERDAS IMAGINE are widely used for georeferencing tasks.

• Integration of automated algorithms and AI is increasingly reducing human effort and improving precision.

?? Challenges in Georeferencing

• Image Distortion: Caused by terrain, atmospheric conditions, or sensor orientation.

• Control Point Selection: Accuracy depends on number, distribution, and quality of Ground Control Points (GCPs).

• Image Resolution & Scale: High-resolution images need precise control points; low-resolution images may lack sufficient detail for accurate georeferencing.

???? Literature Review: Evolution of Georeferencing Techniques

• Early Techniques:

  • Smith et al. (1995): Manual methods using affine transformations.

  • Jones & Brown (1998): Polynomial transformations improved modeling of distortions.

• GIS Integration & Historical Maps:

  • Yuan & Elaksher (2007): Addressed image distortions using GIS techniques.

  • Rumsey & Williams (2002): Studied distortion impacts in historical maps.

• Advancements in Automation:

  • Chiabrando et al. (2016): Used machine learning and crowdsourcing for georeferencing.

  • Zhang & Luo (2018): Proposed real-time georeferencing for UAVs.

  • Campbell & Wilke (2020): Integrated OpenStreetMap (OSM) data for enhanced automation.

  • Johnson & O'Reilly (2021): CNNs significantly improved performance in remote sensing images.

• Hybrid Methods:

  • Liu et al. (2018): Combined pixel- and feature-based techniques for better efficiency and accuracy.

  • Liu & Chen (2020): Showed hybrid methods were ideal for high-resolution images.

???? Comparative Studies & Evaluations

• Martinez & Garcia (2017): Feature-based methods (SIFT, SURF) more accurate than affine/polynomial methods.

• Smith & Brown (2018): Machine learning methods outperformed traditional ones in remote sensing.

• Johnson & Wang (2019): Automated methods were faster; manual methods better with fewer GCPs.

• Lee & Park (2021): OSM-based algorithms improved accuracy, especially when using machine learning.

???? Composite Georeferencing Method (CGM)

• CGM integrates both control points (CP) and check points (Ch.P) for transformation accuracy.

• Uses affine or Helmert transformations, calculating distances between points for stepwise adjustments.

Conclusion

Figure 18 presents a summary of the final results, showing that all experiments yield better outcomes compared to the traditional Helmert and Affine methods. Notably, it demonstrates that the fifth-degree Cosine Influence Factor is the optimal choice for applying CGM. In Figure 19, it is shown that for about the first quarter of the maximum distance in the image, the cosine and Tangent Influence Factors behave similarly, whereas the linear Influence Factor has a weaker effect.

References

[1] Angela M. Peters and Robert J. Smith, 2022. \\\"The Use of Control Points in Image Georeferencing: A Critical Review\\\" [2] Chiabrando, F., Donadio, E., & Rinaudo, F, 2016. \\\"Automated Georeferencing of Historical Maps Using Machine Learning and Crowdsourcing Techniques\\\", ISPRS International Journal of Geo-Information. [3] Ian H. Campbell and Naomi L. Wilke, 2020. \\\"Automated Image Georeferencing Using OpenStreetMap Data\\\" [4] Jennifer L. McDaid and David W. Long, 2019. \\\"A Comparative Study of Georeferencing Algorithms for Historical Maps\\\" [5] Johnson, M., & Wang, Y, 2019. \\\"Performance Comparison of Automated and Manual Georeferencing Methods for High-Resolution Satellite Images\\\" [6] Lee, H., & Park, S., 2021. \\\"A Comparative Study of Georeferencing Algorithms Using OpenStreetMap Data\\\" [7] Li, D., Wang, J., & Zhang, X., 2012. \\\"Georeferencing of Satellite Images Using Ground Control Points and Feature Matching Techniques\\\", Photogrammetric Engineering & Remote Sensing. [8] Liu, X., & Chen, G., 2020. \\\"Hybrid Georeferencing Approaches: A Comparative Study\\\" [9] Liu, X., Zhang, Y., & Chen, G., 2018. \\\"Hybrid Georeferencing Method for High-Resolution Satellite Imagery\\\" [10] Maria B. Johnson and Kevin R. O\\\'Reilly, 2021. \\\"Georeferencing of Remote Sensing Images with Convolutional Neural Networks\\\" [11] Martinez, J., & Garcia, L., 2017. \\\"Comparison of Georeferencing Methods for Historical Map Rectification\\\" [12] Rumsey, D., & Williams, M., 2002. \\\"Georeferencing Historical Maps in a GIS Environment\\\", Cartography and Geographic Information Science. [13] Smith, R., & Brown, T., 2018. \\\"Evaluating the Accuracy of Image Georeferencing Techniques in Remote Sensing\\\" [14] Yuan, F., & Elaksher, A., 2007. \\\"Georeferencing of Historical Aerial Photographs Using GIS and Image Processing Techniques\\\", IEEE Geoscience and Remote Sensing Letters. [15] Zhang, L., & Luo, H., 2018. \\\"Real-time Image Georeferencing Using UAV-Based Remote Sensing\\\", Remote Sensing.

Copyright

Copyright © 2025 Hasan M. Bilani, Ismat Elhasan, Maan. Okayli, Abdullah M. Al-Enazy, Ahmad H. Alashaikh. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.