Abstract:Aircraft recognition is crucial in both civil and military fields, and high-spatial resolution remote sensing has emerged as a practical approach. However, existing data-driven methods fail to locate discriminative regions for effective feature extraction due to limited training data, leading to poor recognition performance. To address this issue, we propose a knowledge-driven deep learning method called the explicable aircraft recognition framework based on a part parsing prior (APPEAR). APPEAR explicitly models the aircraft's rigid structure as a pixel-level part parsing prior, dividing it into five parts: 1) the nose; 2) left wing; 3) right wing; 4) fuselage; and 5) tail. This fine-grained prior provides reliable part locations to delineate aircraft architecture and imposes spatial constraints among the parts, effectively reducing the search space for model optimization and identifying subtle interclass differences. A knowledge-driven aircraft part attention (KAPA) module uses this prior to achieving a geometric-invariant representation for identifying discriminative features. Part features are generated by part indexing in a specific order and sequentially embedded into a compact space to obtain a fixed-length representation for each part, invariant to aircraft orientation and scale. The part attention module then takes the embedded part features, adaptively reweights their importance to identify discriminative parts, and aggregates them for recognition. The proposed APPEAR framework is evaluated on two aircraft recognition datasets and achieves superior performance. Moreover, experiments with few-shot learning methods demonstrate the robustness of our framework in different tasks. Ablation analysis illustrates that the fuselage and wings of the aircraft are the most effective parts for recognition.

Aircraft Recognition Based on Improved Iterative Threshold Selection and Skeleton Zernike Moment

Aircraft Recognition in Low-Resolution Sar Imagery Using Peak Feature Matching

Remote Aircraft Target Recognition Method Based on Superpixel Segmentation and Image Reconstruction

Research on Iterative Segmentation Algorithm Based on Moment and Adaptive Threshold

A Novel Method of Aircraft Detection under Complex Background Based on Circular Intensity Filter and Rotation Invariant Feature

Research on Aerial Photography Target Detection Methods

Aircraft Recognition Based on Mixed Moment Algorithm

Automatic Target Recognition Method Based on Sequential Images

Aircraft Target Detection in Remote Sensing Images Based on Improved YOLOv5

A New Method for Identifying 3D Aircraft Targets

Aircraft Target Detection in Low Signal-to-Noise Ratio Visible Remote Sensing Images

Aircraft Target Detection from Remote Sensing Images under Complex Meteorological Conditions

Method of Target Recognition from Low SNR Building FLIR Images Based on Template

Aircraft Recognition in Remote Sensing Images Based on Deep Learning

Aircraft remote sensing image target detection based on shallow feature enhancement and k-means clustering algorithm

An Aircraft Extraction Method of High-resolution Imagery Based on Shape Feature

Fast target recognition algorithm based on Pseudo-Zernike moment and support vector machine

A Method of Target Recognition from Remote Sensing Images

Explicable Fine-Grained Aircraft Recognition Via Deep Part Parsing Prior Framework for High-Resolution Remote Sensing Imagery

A METHOD OF AIRCRAFT TARGET RECOGNITION BASED ON LLE AND HMM