Huasheng Huang

In the field of precision agriculture, accurately detecting rice panicles is crucial for monitoring rice growth and managing rice production. To address the challenges posed by complex field backgrounds, including variety differences, variations across growth stages, background interference, and occlusion due to dense distribution, this study develops an improved YOLO11-based rice panicle detection model, termed DRPU-YOLO11. The model incorporates a task-oriented CSP-PGMA module in the backbone to enhance multi-scale feature extraction and provide richer representations for downstream detection. In the neck network, DySample and CGDown are adopted to strengthen global contextual feature aggregation and suppress background interference for small targets. Furthermore, fine-grained P2 level information is integrated with higher-level features through a cross-scale fusion module (CSP-ONMK) to improve detection robustness in dense and occluded scenes. In addition, the PowerTAL strategy adapts quality-aware label assignment to emphasize high-quality predictions during training. The experimental results based on a self-constructed dataset demonstrate that DRPU-YOLO11 significantly outperforms baseline models in rice panicle detection under complex field environments, achieving an accuracy of 82.5%. Compared with the baseline model YOLO11 and RT-DETR, the mAP₅₀ increases by 2.4% and 5.0%, respectively. These results indicate that the proposed task-driven design provides a practical and high-precision solution for rice panicle detection, with potential applications in rice growth monitoring and yield estimation.

Nowadays, the combination of UAV remote sensing and deep learning has facilitated effective high-throughput field phenotyping for rice blast-resistant breeding. However, breeding practices can hardly provide sufficient samples for each category due to the the limitations of germplasm resources, which may cause the data insufficiency and class imbalance. In addition, foliar and lesion details are often difficult to identify in UAV images, due to the limitation of spatial resolution. As a result, the application of deep learning can lead to overfitting, as the model may struggle to acquire discriminative features. While previous studies have attempted to combine handcrafted and deep features to address problems with data insufficiency and class imbalances, image degradation still prevents the network from learning efficient representations for disease identification. To address these issues, this paper proposes a hybrid integrated feature fusion (HIFF) method, in which a novel handcrafted design guided CNN (HDGC) module was employed to alleviate the problem of image degradation. Both handcrafted and deep learning branches were integrated in an end-to-end structure and applied to rice blast resistance identification. The proposed method was carefully evaluated using an ablation study, and the comparisons with state-of-the-art deep learning and feature fusion methods were conducted to demonstrate its superiority. Experimental results showed that the HIFF model outperformed mainstream methods by 0.0353 in F1-score and 0.0488 in accuracy on the practical rice blast-resistant breeding applications. As such, the proposed method could be used to accelerate the process of rice blast-resistant breeding. Our implementations are available at: http://github.com/huanghsheng/HIFF.

Raman spectroscopy is a spectral analysis technique based on molecular vibration. It has gained widespread acceptance as a practical tool for the non-invasive and rapid characterization or identification of multiple analytes and compounds in recent years. In fruit quality detection, Raman spectroscopy is employed to detect organic compounds, such as pigments, phenols, and sugars, as well as to analyze the molecular structures of specific chemical bonds or functional groups, providing valuable insights into fruit disease detection, pesticide residue analysis, and origin identification. Consequently, Raman spectroscopy techniques have demonstrated significant potential in agri-food analysis across various domains. Notably, the frontier of Raman spectroscopy is experiencing a surge in machine learning applications to enhance the resolution and quality of the resulting spectra. This paper reviews the fundamental principles and recent advancements in Raman spectroscopy and explores data processing techniques that use machine learning in Raman spectroscopy, with a focus on its applications in detecting fruit diseases, analyzing pesticide residues, and identifying origins. Finally, it highlights the challenges and future prospects of Raman spectroscopy, offering an effective reference for fruit quality detection.

Color calibration is a critical step for Unmanned Aerial Vehicles (UAV) remote sensing, especially in precision agriculture, which relies mainly on correlating color changes to specific quality attributes, e.g., plant health, disease, and pest stresses. In UAV remote sensing, the exemplar-based color transfer is popularly used for color calibration, where the automatic search for the semantic correspondences is the key to ensuring the color transfer accuracy. However, the existing attention mechanisms encounter difficulties in building the precise semantic correspondences between the reference image and the target one, in which the normalized cross correlation is often computed for feature reassembling. As a result, the color transfer accuracy is inevitably decreased by the disturbance from the semantically unrelated pixels, leading to semantic mismatch due to the absence of semantic correspondences. In this paper, we proposed an unsupervised object-based attention mechanism (OBAM) to suppress the disturbance of the semantically unrelated pixels, along with a further introduced weight-adjusted AdaIN (WAA) method to tackle the challenges caused by the absence of semantic correspondences. By embedding the proposed modules into a photorealistic style transfer method with progressive stylization, the color transfer accuracy can be improved while better preserving the structural details. We evaluated our approach on the UAV data of different crop types including rice, beans, and cotton. Extensive experiments demonstrate that our proposed method outperforms several state-of-the-art methods. As our approach requires no annotated labels, it can be easily embedded into the off-the-shelf color transfer approaches. Relevant codes and configurations will be available at http://github.com/huanghsheng/object-based-attention-mechanism.