The field of sustainable and intelligent phytoprotection is witnessing rapid progress at the intersection of plant science, computer vision, and robotics. Effective phytoprotection relies on early and accurate detection of plant rows and their associated characteristics, as this underpins precision management and the implementation of targeted agronomic interventions. Autonomous crop row detection, supported by advances in visual perception technologies, is proving crucial for both ground-based robots and UAVs tasked with monitoring field performance, identifying stress symptoms, and assessing planting quality. UAV-based detection not only enables large-scale, rapid assessment of uniformity, emergence, and growth dynamics but also offers potential for early diagnosis of biotic and abiotic stresses, pest outbreaks, and spatial variability in plant health—essential factors for protecting crop yields and ecosystem stability.

While traditional image processing approaches have long been used, they often fail to capture the biological nuances of plant architecture, canopy density, and intra-row variability under fluctuating field conditions, such as weed competition, disease symptoms, and variable illumination. Recent studies leveraging deep learning and multimodal data (e.g., RGB, multispectral, and LiDAR) demonstrate improved detection and classification, yet gaps persist. For instance, the translation of robust laboratory results to dynamic field environments is hindered by biological complexity—such as overlapping canopies, weed-crop spectral similarity, and differing crop phenologies—that challenge generalizability and accuracy. Despite progress in scaling up sensing and algorithms, the biological underpinnings of plant perception and stress signatures often remain underexplored, limiting the comprehensive integration of phenotyping, pathology, and ecological feedbacks in autonomous systems.

This Research Topic aims to catalyze progress in sustainable and intelligent phytoprotection by advancing robust, biologically informed visual perception for autonomous crop row detection. We seek contributions that not only improve detection accuracy, efficiency, and adaptability across diverse field conditions, but also explicitly integrate plant physiological, pathological, and ecological knowledge into perception workflows. Our overarching goal is to bridge technical and biological domains to foster real-time and resilient phytoprotection—enabling timely interventions against pests, diseases, and stresses, while underpinning sustainable yield optimization. We invite research extending beyond controlled environments, addressing real-world complexity and variability, and integrating multidisciplinary insights from plant science, agronomy, remote sensing, and machine learning. By uniting technological and biological innovation, this Research Topic will chart new directions for holistic and sustainable crop production systems.

The scope of this Research Topic centers on enhancing visual perception technologies for crop row detection with explicit biological integration and phytoprotection impact. Submissions must focus on methods, systems, or field studies that advance understanding or practical application in plant-centric row detection and related crop protection strategies. To gather further insights in sustainable and intelligent phytoprotection, we welcome articles addressing, but not limited to, the following themes:

• Deep learning and computer vision approaches for plant-oriented row identification and stress detection;

• Multimodal data integration (RGB, multispectral, hyperspectral, LiDAR) for distinguishing crop, weed, and disease signatures;

• Domain adaptation for row detection across crops of varying phenology, growth stages, and stress environments;

• Real-time and resource-efficient crop protection algorithms deployable on agricultural robots and UAVs;

• Biological interpretation of image-based features relevant to plant health, disease, pest, and nutrient status;

• Case studies on the impact of improved row detection on disease management, weed suppression, and sustainable yield;

• Field validation of intelligent perception systems aligning with agronomic and ecological outcomes.

A variety of manuscript types are encouraged, including original research articles, reviews, methods papers, and technology reports, especially those reflecting collaborations between plant scientists, agronomists, remote sensing specialists, computer vision experts, and agricultural engineers.

Are you interested in contributing a manuscript to this Research Topic?

Advancements and Challenges in Visual Perception for Autonomous Crop Row Detection in Agriculture