Cotton is the most important natural fiber crop worldwide, and its automated harvesting is essential for improving production efficiency and economic benefits. However, cotton boll detection faces challenges such as small target size, fine-grained category differences, and complex background interference. This study proposes RA-CottNet, a high-precision object detection model with both directional awareness and attention-guided capabilities, and develops an open-source dataset containing 4966 annotated images. Based on YOLOv11n, RA-CottNet incorporates ODConv and SPDConv to enhance directional and spatial representation, while integrating CoordAttention, an improved GAM, and LSKA to improve feature extraction. Experimental results showed that RA-CottNet achieves 93.683% $Precision$, 86.040% $Recall$, 93.496% $mAP50$, 72.857% $mAP95$, and 89.692% $F1$-$score$, maintaining stable performance under multi-scale and rotation perturbations. The proposed approach demonstrated high accuracy and real-time capability, making it suitable for deployment on agricultural edge devices and providing effective technical support for automated cotton boll harvesting and yield estimation.

Accurate, high-throughput quantification of insect-induced crop injury is critical for plant breeding and crop management, but traditional visual assessments are labor-intensive and subjective. This study evaluated unmanned aerial vehicle (UAV)-derived imagery to quantify cotton jassid injury severity across multiple cotton populations and timepoints. A total of 241 predictors were generated from combinations of vegetation indices, image segmentation methods, and summary statistics derived from UAV imagery. Correlation analysis identified metrics derived from the Excess Red (ExR) and Excess Red minus Green (ExGR) indices as the strongest predictors of injury severity, with correlation coefficients up to 0.64. Segmentation improved associations with manual ratings relative to unsegmented imagery, with IsoCluster-derived predictors achieving the strongest correlations. Quantitative evaluation of segmentation accuracy demonstrated that the Segment Anything Model v3 (SAM3) provided the most accurate canopy delineation, highlighting differences between segmentation fidelity and downstream trait prediction. Multivariate regression models integrating key predictors achieved substantial agreement with manual ratings (Pearson r = 0.75, Root Mean Square Error ($RMSE$) = 0.79), while a mixed-effects model accounting for genotype and field variation further improved in-sample agreement (r = 0.82, $RMSE$ = 0.69). Inclusion of a nonlinear canopy area term improved model fit, reflecting the threshold response of canopy structure to increasing injury severity. Resulting models, combined with manual ratings, identified several candidate lines exhibiting consistently reduced cotton jassid-associated injury. These results demonstrate that UAV-based phenotyping provides a scalable, objective approach for quantifying cotton jassid injury and supports high-throughput evaluation of cotton germplasm to identify reduced-injury candidates for subsequent resistance evaluation.

Areolate mildew is a foliar disease of cotton, with two fungal species reported as causal pathogens. We announce the first genome of Ramulariopsis gossypii . The genome size is 33.34 Mbp, with an N50 of 2.116 Mb and Benchmarking Universal Single-Copy Orthologs completeness of 98.8%, providing a resource for comparative genomics and disease management.