Description:FIELD OF THE INVENTION
The present invention relates to the field of artificial intelligence and satellite image analysis. More specifically, it pertains to an explainable AI-based hybrid classification system using Logical Neural Networks and Q-Lattice for interpretable and accurate classification of multispectral satellite images.
BACKGROUND OF THE INVENTION
Black box testing produces a problem because its inner decision systems remain opaque to analysis. The insufficient interpretability of these deep learning models creates substantial difficulties while they process satellite images for classification purposes and nevertheless makes their implementation problematic when applications need to understand decision-making processes in land cover monitoring and disaster response alongside environmental analysis.
Modern remote sensing spectral band selection methods fail to detect trustworthy wavelengths that would lead to successful classification operations. Users become confused with AI predictions because they receive no clear explanations about them. Scientific research adopting XAI systems needs accurate satellite photo categorization functions along with detailed explanations of prediction outcomes. AI explanation systems provide businesses with a framework to use remote sensing systems based on AI technology safely.
No such model exists. The current analytical models implement CNN features together with ViT features and Grad-CAM and SHAP methods yet these methods do not meet the requirements for comprehensive spectral analysis in classification tasks. The system requires identifying crucial spectral bands together with complete rationalization behind decision ranking procedures.
The designed architecture provides more benefits than standard models and improve diagnostic accuracy as well as simplifies interpretation. Structured reasoning combined with logical restrictions is implemented in Logical Neural Networks (LNNs) which are also more interpretable due to the fact that their decision-making processes ameliorate the procedures. By identifying and ranking vital features in Q-Lattice, spectral feature analysis results can be much better than \ without having to duplicate. With this hybrid approach, the performance is superior in terms of efficiency with the ability to be deployed at the cost of expensive processing requirements of Vision Transformers (ViTs) and CNNs. Reliable forecasting outcomes can be generated by the LNN-Q-Lattice system to instill trust in AI decisions in environmental surveillance tasks, disaster mitigation operations as well as in land resource evaluation exercises. Through the system the users gain visual and verbal explanations that show them why the model took the decisions it made and that increases their confidence. Models are made to learn from a variety of datasets from satellites, practically regardless of dataset, so they adapt to unseen information outside of EuroSat. A solution framework is proposed to solve deep learning interpretability problems with the purpose of providing transparent satellite image classification solutions.
OBJECTIVES OF THE INVENTION
Main objective of the present invention is to develop a hybrid explainable artificial intelligence (XAI) system combining Logical Neural Networks (LNNs) and Q-Lattice for accurate and interpretable satellite image classification.
Another objective of the present invention is to utilize logic-based constraints within LNNs to embed domain knowledge and enhance the transparency of classification decisions.
Another objective of the present invention is to implement Q-Lattice explainability techniques for effective feature selection and identification of significant spectral bands in multispectral satellite images.
Another objective of the present invention is to generate both visual and textual explanations accompanying classification outputs for improved interpretability and trust in AI systems.
Another objective of the present invention is to evaluate and validate the proposed framework using metrics such as accuracy, F1-score, and interpretability, and demonstrate its application in emergency response and environmental monitoring tasks.
SUMMARY OF THE INVENTION
This summary is provided to introduce a selection of concepts, in a simplified format, that are further described in the detailed description of the invention.
This summary is neither intended to identify key or essential inventive concepts of the invention and nor is it intended for determining the scope of the invention.
To further clarify advantages and features of the present invention, a more particular description of the invention will be rendered by reference to specific embodiments thereof, which is illustrated in the appended drawings. It is appreciated that these drawings depict only typical embodiments of the invention and are therefore not to be considered limiting of its scope. The invention will be described and explained with additional specificity and detail with the accompanying drawings.
The proposed hybrid system combines Logical Neural Networks (LNNs) unification with Q-Lattice Explainable AI (XAI) to address existing issues and enhance predictive performance and accurate classification. The system uses optimally normalized transformation techniques to analyze multispectral satellite pictures that come from the EuroSAT dataset. LNNs implement systematic decisions from beginning to end because of their logic-based restrictions which embed domain expertise to establish clear explanations. Q-Lattice achieves explainability through its assessment of feature relationships to identify significant spectral bands used in categorization which results in visual explanations that accompany verbal descriptions for improved transparency. The validation process showed success through results that compared LNNs with ResNeXt-50 (32x4d) and Wide ResNet-50-2 models under accuracy, F1-score measures and interpretability analysis. The fusion algorithm achieves high-level accuracy performance and offers easy-to-understand logical explanations which enable its application for emergency response operations together with environmental surveillance systems.
BRIEF DESCRIPTION OF THE DRAWINGS
The illustrated embodiments of the subject matter will be understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of devices, systems, and methods that are consistent with the subject matter as claimed herein, wherein:
FIGURE 1: BLOCK DIAGRAM OF THE PROPOSED SYSTEM
The figures depict embodiments of the present subject matter for the purposes of illustration only. A person skilled in the art will easily recognize from the following description that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles of the disclosure described herein.
DETAILED DESCRIPTION OF THE INVENTION
The detailed description of various exemplary embodiments of the disclosure is described herein with reference to the accompanying drawings. It should be noted that the embodiments are described herein in such details as to clearly communicate the disclosure. However, the amount of details provided herein is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the present disclosure as defined by the appended claims.
It is also to be understood that various arrangements may be devised that, although not explicitly described or shown herein, embody the principles of the present disclosure. Moreover, all statements herein reciting principles, aspects, and embodiments of the present disclosure, as well as specific examples, are intended to encompass equivalents thereof.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments. As used herein, the singular forms “a",” “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises,” “comprising,” “includes” and/or “including,” when used herein, specify the presence of stated features, integers, steps, operations, elements and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof.
It should also be noted that in some alternative implementations, the functions/acts noted may occur out of the order noted in the figures. For example, two figures shown in succession may, in fact, be executed concurrently or may sometimes be executed in the reverse order, depending upon the functionality/acts involved.
In addition, the descriptions of "first", "second", “third”, and the like in the present invention are used for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Thus, features defining "first" and "second" may include at least one of the features, either explicitly or implicitly.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which example embodiments belong. It will be further understood that terms, e.g., those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
In some embodiments of the present invention, the present invention discloses a hybrid explainable artificial intelligence (XAI) framework for satellite image classification that integrates Logical Neural Networks (LNNs) with a Q-Lattice explainability module.
In some embodiments of the present invention, the system processes multispectral images from the EuroSAT dataset through a preprocessing module involving normalization, transformation, and feature extraction.
In some embodiments of the present invention, the LNNs perform logic-constrained classification embedding domain knowledge to produce interpretable decisions, while the Q-Lattice module identifies significant spectral bands and elucidates feature relationships.
In some embodiments of the present invention, the framework outputs both visual and textual explanations to enhance model transparency. Evaluation results demonstrate superior performance in accuracy, F1-score, and interpretability compared to conventional deep learning models such as ResNeXt-50 (32x4d) and Wide ResNet-50-2.
In some embodiments of the present invention, this invention addresses the black-box limitations of traditional neural networks and offers an interpretable and accurate solution for critical applications such as emergency response and environmental monitoring. It further represents the first application of LNNs in satellite image classification.
Herein enclosed a hybrid explainable artificial intelligence (XAI) system for satellite image classification, comprising:
an input layer configured to receive multispectral satellite images from the EuroSAT dataset;
a preprocessing module comprising normalization, transformation, and feature extraction units for processing the input images;
a classification model comprising a hybrid architecture integrating Logical Neural Networks (LNNs) and a Q-Lattice explainability module;
wherein the Logical Neural Networks (LNNs) perform logic-constrained decision-making to enable transparent and explainable classification;
and the Q-Lattice explainability module identifies significant spectral features and relationships to generate interpretable outputs;
the system further configured to output visual and textual explanations corresponding to the classification results.
The LNNs embed domain knowledge via logic-based restrictions to provide clear and systematic decision pathways.
The Q-Lattice explainability module performs feature selection and model explanation by evaluating the relevance and interaction of spectral bands.
The preprocessing module applies optimal normalization and transformation techniques to improve feature relevance prior to classification.
The evaluation metrics include accuracy, F1-score, and interpretability scores to validate classification and explainability performance.
EXAMPLE 1
BEST METHOD
As a proposed plan and architecture, we have developed a Pseudocode for a Hybrid Liquid Neural Network (LNN) and Q-Lattice XAI Framework for satellite image classification. The envisioned architecture utilizes LNNs for the classification and Q-Lattice Explainable AI for feature selection and model explanation.
The proposed hybrid system combines Logical Neural Networks (LNNs) unification with Q-Lattice Explainable AI (XAI) to address existing issues and enhance predictive performance and accurate classification. The system uses optimally normalized transformation techniques to analyze multispectral satellite pictures that come from the EuroSAT dataset. LNNs implement systematic decisions from beginning to end because of their logic-based restrictions which embed domain expertise to establish clear explanations. Q-Lattice achieves explainability through its assessment of feature relationships to identify significant spectral bands used in categorization which results in visual explanations that accompany verbal descriptions for improved transparency. The validation process showed success through results that compared LNNs with ResNeXt-50 (32x4d) and Wide ResNet-50-2 models under accuracy, F1-score measures and interpretability analysis. The fusion algorithm achieves high-level accuracy performance and offers easy-to-understand logical explanations which enable its application for emergency response operations together with environmental surveillance systems.
The new approach presents a transparent AI framework that resolves deep learning black-box issues to facilitate logical decision making by essential applications. First application of LNNs for satellite image classification.
NOVELTY:
• Integration of Q-Lattice for enhanced feature importance analysis.
• Improved interpretability in remote sensing AI models compared to existing black-box approaches.
The proposed hybrid framework combines Logical Neural Networks (LNNs) alongside Q-Lattice Explainable AI (XAI) to achieve improved classification precision while maintaining explainable satellite image analysis capabilities for structured decision systems and feature importance assessments. 
, Claims:1. A hybrid explainable artificial intelligence (XAI) system for satellite image classification, comprising:
an input layer configured to receive multispectral satellite images from the EuroSAT dataset;
a preprocessing module comprising normalization, transformation, and feature extraction units for processing the input images;
a classification model comprising a hybrid architecture integrating Logical Neural Networks (LNNs) and a Q-Lattice explainability module;
wherein the Logical Neural Networks (LNNs) perform logic-constrained decision-making to enable transparent and explainable classification;
and the Q-Lattice explainability module identifies significant spectral features and relationships to generate interpretable outputs;
the system further configured to output visual and textual explanations corresponding to the classification results.
2. The system as claimed in claim 1, wherein the LNNs embed domain knowledge via logic-based restrictions to provide clear and systematic decision pathways.
3. The system as claimed in claim 1, wherein the Q-Lattice explainability module performs feature selection and model explanation by evaluating the relevance and interaction of spectral bands.
4. The system as claimed in claim 1, wherein the preprocessing module applies optimal normalization and transformation techniques to improve feature relevance prior to classification.
5. The system as claimed in claim 1, wherein evaluation metrics include accuracy, F1-score, and interpretability scores to validate classification and explainability performance.