What can SuperMap GeoAI do? (Ⅱ)

SuperMap AI Remote Sensing Image Processing

Deep Integration with AI Technology, Providing Intelligent Processing Capabilities

1.AI Image Matching Technology

1.1Enhancing Heterogeneous Image Matching Accuracy

?Traditional matching algorithms: Limited number of matching points, lower matching accuracy

?AI Matching Algorithms: Significant increase in the number of matching points, higher matching accuracy

*Compared to traditional algorithms, AI algorithms can address matching challenges such as significant lighting differences, large contrast variations, and non-linear radiation distortions in heterogeneous images, therefore enabling robust matching of heterogeneous images.        

1.2 Robust Matching of Weak-Texture Images

*The screenshot uses the MOTIF matching algorithm, and during the matching process, a single band of the reference image is used, resulting in a grayscale effect.        

2. AI Semantic-Assisted Geometric Processing

2.1 Improving Geometric Processing Accuracy

• A comparison of traditional geometric processing and AI semantic-assisted geometric processing

3. AI DSM Algorithm

3.1 Enhancing Integrity and Smoothness in Texture-Deficient Areas

• Traditional Algorithms: Presence of holes and noise

• AI Algorithms: No holes, better completeness

3.2 Improving Surface Smoothness and Feature Recognition

• A comparison of traditional methods and AI methodes

4. AI Enables Efficient Latte Detection

• Mountain Latte Detection

• Building/Road Latte Detection

*Background knowledge
Two major causes of image latte:
a. DEM Data Errors: Significant deviation from real terrain
b. The object coordinates of adjacent pixels are close yet with large elevation differences        

5. AI Automated Cloud Removal in Remote Sensing Images

6. AI Automated Repair of DSM/DEM Water Surfaces

• Abnormal Water Elevation Values

• Repaired Water Surface Effects

Scenario 1：Water Data Provided
Elevation information included in data attributes, realizing automated value assignment        

Scenario 2：No Water Data Provided
a. Remote sensing AI extracts water surface from original Images
b. Automated water elevation supplement based on reference DEM
c. Water surface repair through combining Various Information        

7.AI-Based Automated Quality Inspection of Orthophotos, Improving Manual Inspection Efficiency and Accuracy

• Hierarchical display of inspection results significantly improve manual inspection efficiency

• AI automated quality inspection of fine grid partition, enhancing inspection accuracy

*Traditional accuracy reports are general and not intuitive, requiring screen-by-screen scrolling which is time-consuming;
SuperMap production results not only provide inspection reports but also display errors hierarchically based on their magnitude;
Users only nees to focus on red/orange point-dense areas to quickly locate high-risk regions, which will improve inspection efficiency;
For example, in the image in the last page, the traditional methods require 54 screens for scrolling, but now only 5 areas need attention.        

Coming next: AI Remote Sensing Image Interpretation