Global climate change and intensive human activities have led to increasing degradation of arable land due to ionic stress, such as salinization and heavy metal contamination, posing a huge threat to world food security. The rapid development of artificial intelligence (AI) technology offers a new approach for germplasm innovation for stress adaptation. This review systematically summarizes key advances and applications of AI in germplasm innovation for ionic stress tolerance. In the field of phenomics, multi-dimensional phenotyping systems integrate technologies including red-green-blue imaging, spectral sensing, fluorescence imaging, thermal infrared imaging, and X-ray fluorescence spectrometer. These are coupled with machine learning, deep learning, and multi-source data fusion to enable early, accurate, and non-invasive identification of ionic stress responses. In the field of smart breeding, machine learning accelerates gene discovery, along with pangenomics to elucidate genetic diversity. AI-driven prediction of multi-gene regulatory networks and genomic selection have significantly improved breeding efficiency. However, challenges such as insufficient data standardization and sharing, limited prediction accuracy of genotype-environment interaction (G×E) models, and high technology promotion costs are still constraining the large-scale application of the AI-enabled technologies. We expect that future efforts with focus on data integration, algorithm optimization, and technology popularization will allow AI to play a greater role in the remediation of degraded farmland and in achieving sustainable agricultural development.

Focusing on the core issue of crop phenotyping empowered by artificial intelligence (AI), this paper systematically summarizes the typical biological characteristics of various crop phenotypes and the corresponding high-throughput phenotyping acquisition techniques from six dimensions, namely morphological structure, growth and development, physiology and biochemistry, yield and quality, stress and disease resistance, and molecular genetics. Aiming at the above six dimensions, it emphatically reviews the latest research findings and application progress of AI technologies in the automatic identification of plant phenotypic characteristics, the accurate extraction of multi-scale phenotypic parameters, and the modeling and analysis of multi-source phenotypic data. Finally, it analyzes the key challenges and deficiencies faced by current AI-driven crop phenotyping research in terms of adaptability to complex field environments, fusion of multi-source heterogeneous data, and accurate genotype-to-phenotype (G2P) modeling. Furthermore, it prospects the future development trends of AI technologies in the fields of crop phenotyping acquisition, analysis and phenomics, as well as their application prospects and research directions in intelligent breeding and agricultural intelligent decision-making.

With global warming and population growth accelerating, crop breeding must advance towards greater precision, efficiency and sustainability. Riding the wave of rapid AI development, AI-driven smart breeding has emerged as the cutting edge of plant breeding. Leveraging technologies such as high-throughput phenotyping through machine learning and deep learning, genome-wide association studies for genomic selection, and multiomics big data analysis, it integrates and edits genetic information while linking phenotypes with genomes. This approach promises revolutionary changes in agricultural breeding, enhancing crop productivity, improving quality, and refining traits. Looking ahead, precision agriculture and personalised breeding represent key developmental trajectories. Strategies focusing on environmental adaptability and climate change mitigation will aid in addressing ecological challenges. The cutting-edge technologies of smart breeding, significantly enhancing breeding efficiency and precision, will shape the future direction and progress of plant breeding.

Smart agriculture, as a core direction of modern agricultural development, is driving the transformation of agricultural production from experience-driven to data-driven, and from manual operations to intelligent and automated systems. This paper systematically reviews the applications of artificial intelligence (AI) in smart agriculture and the current status, trends, and challenges of smart cultivation, constructing a “sensing-decision-execution-management” closed-loop framework to elucidate its role and value across the entire agricultural process. The sensing layer leverages multi-source sensors, imaging technologies, and the Internet of Things (IoT) to achieve precise monitoring of crops and environmental conditions; the decision layer integrates data-driven algorithms and crop growth models to enable intelligent prediction and optimization of fertilization, irrigation, and pest and disease management; the execution layer employs unmanned agricultural machinery, drones, and robots to perform high-precision autonomous operations; and the management layer utilizes cloud-edge collaboration and digital twins to realize visualized and sustainable farm management. Although smart agriculture is developing rapidly, challenges remain in data assimilation and standardization, and the level of equipment intelligence and environmental adaptability still needs improvement. Moreover, with the massive acquisition and utilization of agricultural data, issues of data security and privacy protection have become increasingly prominent, posing significant challenges to the sustainable development of smart agriculture. The paper highlights that the core goal of smart cultivation is to increase yield, improve quality, reduce costs, and promote green development, while further advancing multi-source sensing integration, intelligent decision-making loops, autonomous operations, and low-carbon farming. The findings provide a reference for the theoretical framework of smart agriculture and offer technical support for the practical implementation of agricultural IoT, AI, and digital agriculture, playing a significant role in promoting efficient, intelligent, and sustainable agricultural production.

Non-destructive testing of fruit ripeness is of great significance for reducing losses during fruit production, transportation, and storage. Among traditional methods, empirical judgment is highly subjective and less accurate, while destructive testing compromises the integrity of samples. Both methods suffer from low detection efficiency. In recent years, with the rapid development of non-destructive testing technologies such as near-infrared spectroscopy, hyperspectral imaging, machine vision, acoustic vibration, electrical properties, tactile sensing, and electronic noses, the key parameters highly correlated with fruit ripeness-such as color, hardness, sugar-acid ratio, and volatile gases-have been used to accurately assess the maturity state of fruits through machine learning approaches. This study begins with the physiological changes during fruit ripening and systematically reviews the fundamental principles and applications of non-destructive testing technologies. Furthermore, it summarizes current challenges in the field, including environmental interference, lack of data standards, and the balance between cost, complexity, and multimodal integration. The study also explores future directions, such as the development of low-cost multimodal sensor technologies, the establishment of data standards, and improving the interpretability of machine learning.

Horticultural plants are of great significance in cultural inheritance, economic development, and ecosystem sustainability. However, the advancement of horticultural practices is hindered by challenges such as climate change, labor shortages, and inefficient resource utilization. Plant information sensing technology provides innovative solutions to these challenges. Among these technologies, plant wearable sensors have emerged as promising tools for monitoring plant physiological and biochemical information due to their high flexibility, extensibility, spatiotemporal resolution, and biocompatibility, and are becoming essential tools for acquiring data on the plants and their growing environments. Nevertheless, a systematic review of wearable sensor applications in horticultural plant research remains lacking. This review focuses on horticultural plant information detection and comprehensively discusses and summarizes plant wearable sensors’ status of research and applications in the following areas: growth monitoring, moisture content detection, stem sap flow analysis, electrical signal acquisition and chemical substance detection. Finally, we propose future development directions to provide references for information detection in horticultural plants.

Against the dual backdrop of New Agricultural Science development and educational digital transformation, South China Agricultural University addresses core challenges in botany courses at agricultural and forestry universities, such as the disconnection between learning and application, insufficient knowledge internalization, and lack of student interest. With AI as a key enabling tool, the university has constructed a four-in-one support system of ‘resources-tools-scenarios-achievements’ and implemented a closed-loop teaching reform integrating theory-practice, science-education, and industry-education. By reconstructing hierarchical teaching content centered on ‘cell-tissue-organ-group-resource utilization’, building a multi-modal AI intelligent resource matrix, and implementing a ‘three-dimension, three-stage’ process-based assessment, the reform has achieved multi-dimensional improvement of students’ core competencies and qualities, enhanced teachers’ professional capabilities, and highlighted resource radiation effects. This reform provides a replicable paradigm for the digital transformation of botany courses in agricultural and forestry universities, helping to cultivate innovative and compound talents who embody the ‘Four Agriculture’ ethos (understanding agriculture, loving agriculture, strengthening agriculture, and revitalizing agriculture) and serve modern agriculture and rural revitalization.

Guided by the principle of “Empowering Three Dimensions” (intensity, depth, and breadth), this study addresses the major issues in the traditional teaching of plant physiology course, such as fragmented knowledge points, weak connection to practical applications, insufficient student initiatives, and rigid integration of ideological and political education. Taking the course of “the impact of respiration in agricultural and forestry production” as a key case, we explored the integration path of artificial intelligence (AI) and curriculum teaching. By constructing knowledge graphs, designing an AI-driven closed-loop teaching process covering the stages of “pre-class, in-class, and post-class”, and developing case generation tools, our reform has achieved systematic organization of knowledge, personalized teaching, and targeted cultivation of practical abilities of the students. Meanwhile, through AI tools, the course transformed the ideological and political elements—including the wisdom of agricultural civilization and the “Dual carbon” goals—into visual and interactive teaching content, establishing a trinity model of “Knowledge impartment, competence development, and value guidance”. Practice showed that AI empowerment had effectively improved students’ participation and knowledge mastery (e.g., the recognition of carbon balance increased to 89%), enhanced their discipline identity and sense of social responsibility. This research provides practical references for the integrated reform of curriculum intellectualization and ideological and political education.

Under the background of the digital and intelligent transformation in education, knowledge graphs, as a typical representative of new digital technologies, are deeply integrated with teaching and learning, becoming a hotspot in the field of education. Leveraging Xuexitong smart education platform and artificial intelligence (AI) technologies, a systematic knowledge graph for the course of plant biology was constructed in this study, and teaching practices were conducted to explore its substantive role in teaching and learning. The application results indicate that the course knowledge graph not only helps students organize the relationships between learning points, broaden their knowledge horizons, and thus build a systematic knowledge system, but also assists students in formulating personalized learning paths to improve learning abilities. For teachers, the course knowledge graph serves as a powerful tool for efficiently managing teaching resources. Personalized learning data drives teachers to clearly grasp individual student differences, provide precise guidance, and continuously enhance teaching quality. This tripartite collaborative learning ecosystem of teacher-knowledge graph-student provides a highly valuable practical model for the digital and intelligent transformation of education.