1D-RMC: deep hyperspectral sequence modeling for smart remote sensing

José I Cifuentes, Md. Rezwan Parvez, Fernando Castillo, Guillermo Zieballe J, Alejandro Rojas, Hugo O Garcés, Bhushan Gopaluni, Sirish L Shah

Abstract Hyperspectral imaging provides high-dimensional spectral information for material characterization and remote sensing analytics. However, deployment is constrained by hardware cost, inter-band redundancy, and the limited suitability of predominantly spatially oriented models for one-dimensional (1D) spectral signatures. This paper presents an integrated sensor-learning framework that combines data-driven wavelength selection with sequence-aware spectral classification to support compact multispectral sensing. sparse principal component analysis with strong L 1 regularization is used to identify informative wavelengths and derive reduced-band multispectral configurations that mitigate redundancy and the curse of dimensionality. A 1D recurrent multi-head convolutional network is then introduced, which integrates bidirectional recurrent encoding, multi-head self-attention, and hierarchical 1D convolutions to capture local spectral structure and long-range inter-band dependencies. In contrast to convolutional neural network (CNN), recurrent neural network (RNN) and Transformer variants designed for spatial (2D) feature learning or tokenized hyperspectral cubes, the proposed architecture targets reduced-band 1D spectral sequences and is suitable for computationally constrained operation. Experiments on Indian Pines and Houston2013 show that the proposed approach outperforms representative CNN-, RNN-, and transformer-based baselines and requires fewer spectral bands. Overall accuracy reaches 82.3% on Indian Pines and 88.6% on Houston2013.