Crop physiological and nutrient biochemical information plays a vital role in uncovering patterns of crop growth and development, as well as understanding their interactions with environmental factors. Unmanned aerial vehicles (UAV) -based hyperspectral imaging (HSI) technology offers an innovative tool for acquiring physiological and nutrient biochemical information through non-destructive and rapid collection of continuous spectral data from crops. However, challenges such as low signal-to-noise ratios (SNR), spectral variability for the same material, and high dimensionality in hyperspectral data make feature selection and extraction critical steps in data processing and analysis. Therefore, this review focuses on feature selection and extraction methods in the application of UAV-based hyperspectral technology for retrieving and monitoring crop physiological and biochemical information, providing theoretical support for its use in agriculture. Firstly, it provides a detailed discussion of feature selection methods, including filter-based, wrapper-based, and embedded approaches, along with various feature extraction techniques, analyzing their applicability and limitations in crop retrieving and monitoring. Secondly, the review highlights the use of vegetation indices (VIs) in feature extraction, covering advancements from basic indices to those optimized for specific applications. Finally, the article summarizes the main challenges of existing methods, particularly the issues of high-dimensional data processing and noise, and outlines potential future directions. This review highlights the significance of feature selection and extraction methods as critical tools for efficiently processing hyperspectral data. Through systematic analysis and synthesis, it provides theoretical support for agricultural researchers and practitioners while underscoring the importance of these techniques in driving innovation and advancements in hyperspectral technology.

Since the 1950s, when the Turing Test was introduced, there has been notable progress in machine language intelligence. Language modeling, crucial for AI development, has evolved from statistical to neural models over the last two decades. Recently, transformer-based Pre-trained Language Models (PLM) have excelled in Natural Language Processing (NLP) tasks by leveraging large-scale training corpora. Increasing the scale of these models enhances performance significantly, introducing abilities like context learning that smaller models lack. The advancement in Large Language Models, exemplified by the development of ChatGPT, has made significant impacts both academically and industrially, capturing widespread societal interest. This survey provides an overview of the development and prospects from Large Language Models (LLM) to Large Multimodal Models (LMM). It first discusses the contributions and technological advancements of LLMs in the field of natural language processing, especially in text generation and language understanding. Then, it turns to the discussion of LMMs, which integrates various data modalities such as text, images, and sound, demonstrating advanced capabilities in understanding and generating cross-modal content, paving new pathways for the adaptability and flexibility of AI systems. Finally, the survey highlights the prospects of LMMs in terms of technological development and application potential, while also pointing out challenges in data integration, cross-modal understanding accuracy, providing a comprehensive perspective on the latest developments in this field.