Plant Phenomics最新文献

Berry thinning is one of the most important tasks in the management of high-quality table grapes. Farmers often thin the berries per cluster to a standard number by counting. With an aging population, it is hard to find adequate skilled farmers to work during thinning season. It is urgent to design an intelligent berry-thinning machine to avoid exhaustive repetitive labor. A machine vision system that can determine the number of berries removed and locate the berries removed is a challenge for the thinning machine. A method for instance segmentation of berries and berry counting in a single bunch is proposed based on AS-SwinT. In AS-SwinT, Swin Transformer is performed as the backbone to extract the rich characteristics of grape berries. An adaptive feature fusion is introduced to the neck network to sufficiently preserve the underlying features and enhance the detection of small berries. The size of berries in the dataset is statistically analyzed to optimize the anchor scale, and Soft-NMS is used to filter the candidate frames to reduce the missed detection of densely shaded berries. Finally, the proposed method could achieve 65.7 AP^box, 95.0 $A{P}_{0.5}^{\mathrm{box}}$, 57 $A{P}_s^{\mathrm{box}}$, 62.8 AP^mask, 94.3 $A{P}_{0.5}^{\text{mask}}$, 48 $A{P}_s^{\text{mask}}$, which is markedly superior to Mask R-CNN, Mask Scoring R-CNN, and Cascade Mask R-CNN. Linear regressions between predicted numbers and actual numbers are also developed to verify the precision of the proposed model. RMSE and R² values are 7.13 and 0.95, respectively, which are substantially higher than other models, showing the advantage of the AS-SwinT model in berry counting estimation.

The organ-specific critical nitrogen (N_c) dilution curves are widely thought to represent a new approach for crop nitrogen (N) nutrition diagnosis, N management, and crop modeling. The N_c dilution curve can be described by a power function (N_c = A₁·W^-A2), while parameters A₁ and A₂ control the starting point and slope. This study aimed to investigate the uncertainty and drivers of organ-specific curves under different conditions. By using hierarchical Bayesian theory, parameters A₁ and A₂ of the organ-specific N_c dilution curves for wheat were derived and evaluated under 14 different genotype × environment × management (G × E × M) N fertilizer experiments. Our results show that parameters A₁ and A₂ are highly correlated. Although the variation of parameter A₁ was less than that of A₂, the values of both parameters can change significantly in response to G × E × M. Nitrogen nutrition index (NNI) calculated using organ-specific N_c is in general consistent with NNI estimated with overall shoot N_c, indicating that a simple organ-specific N_c dilution curve may be used for wheat N diagnosis to assist N management. However, the significant differences in organ-specific N_c dilution curves across G × E × M conditions imply potential errors in N_c and crop N demand estimated using a general N_c dilution curve in crop models, highlighting a clear need for improvement in N_c calculations in such models. Our results provide new insights into how to improve modeling of crop nitrogen-biomass relations and N management practices under G × E × M.

To better address the difficulties in designing green fruit recognition techniques in machine vision systems, a new fruit detection model is proposed. This model is an optimization of the FCOS (full convolution one-stage object detection) algorithm, incorporating LSC (level scales, spaces, channels) attention blocks in the network structure, and named FCOS-LSC. The method achieves efficient recognition and localization of green fruit images affected by overlapping occlusions, lighting conditions, and capture angles. Specifically, the improved feature extraction network ResNet50 with added deformable convolution is used to fully extract green fruit feature information. The feature pyramid network (FPN) is employed to fully fuse low-level detail information and high-level semantic information in a cross-connected and top-down connected way. Next, the attention mechanisms are added to each of the 3 dimensions of scale, space (including the height and width of the feature map), and channel of the generated multiscale feature map to improve the feature perception capability of the network. Finally, the classification and regression subnetworks of the model are applied to predict the fruit category and bounding box. In the classification branch, a new positive and negative sample selection strategy is applied to better distinguish supervised signals by designing weights in the loss function to achieve more accurate fruit detection. The proposed FCOS-LSC model has 38.65M parameters, 38.72G floating point operations, and mean average precision of 63.0% and 75.2% for detecting green apples and green persimmons, respectively. In summary, FCOS-LSC outperforms the state-of-the-art models in terms of precision and complexity to meet the accurate and efficient requirements of green fruit recognition using intelligent agricultural equipment. Correspondingly, FCOS-LSC can be used to improve the robustness and generalization of the green fruit detection models.

Rapid and accurate estimation of panicle number per unit ground area (PNPA) in winter wheat before heading is crucial to evaluate yield potential and regulate crop growth for increasing the final yield. The accuracies of existing methods were low for estimating PNPA with remotely sensed data acquired before heading since the spectral saturation and background effects were ignored. This study proposed a spectral-textural PNPA sensitive index (SPSI) from unmanned aerial vehicle (UAV) multispectral imagery for reducing the spectral saturation and improving PNPA estimation in winter wheat before heading. The effect of background materials on PNPA estimated by textural indices (TIs) was examined, and the composite index SPSI was constructed by integrating the optimal spectral index (SI) and TI. Subsequently, the performance of SPSI was evaluated in comparison with other indices (SI and TIs). The results demonstrated that green-pixel TIs yielded better performances than all-pixel TIs apart from TI_[HOM], TI_[ENT], and TI_[SEM] among all indices from 8 types of textural features. SPSI, which was calculated by the formula DATT_{[850,730,675]} + NDTI_COR[850,730], exhibited the highest overall accuracies for any date in any dataset in comparison with DATT_{[850,730,675]}, TI_NDRE[MEA], and NDTI_COR[850,730]. For the unified models assembling 2 experimental datasets, the R_V² values of SPSI increased by 0.11 to 0.23, and both RMSE and RRMSE decreased by 16.43% to 38.79% as compared to the suboptimal index on each date. These findings indicated that the SPSI is valuable in reducing the spectral saturation and has great potential to better estimate PNPA using high-resolution satellite imagery.

Evaluation of photosynthetic quantum yield is important for analyzing the phenotype of plants. Chlorophyll a fluorescence (ChlF) has been widely used to estimate plant photosynthesis and its regulatory mechanisms. The ratio of variable to maximum fluorescence, F_v /F_m , obtained from a ChlF induction curve, is commonly used to reflect the maximum photochemical quantum yield of photosystem II (PSII), but it is measured after a sample is dark-adapted for a long time, which limits its practical use. In this research, a least-squares support vector machine (LSSVM) model was developed to explore whether F_v /F_m can be determined from ChlF induction curves measured without dark adaptation. A total of 7,231 samples of 8 different experiments, under diverse conditions, were used to train the LSSVM model. Model evaluation with different samples showed excellent performance in determining F_v /F_m from ChlF signals without dark adaptation. Computation time for each test sample was less than 4 ms. Further, the prediction performance of test dataset was found to be very desirable: a high correlation coefficient (0.762 to 0.974); a low root mean squared error (0.005 to 0.021); and a residual prediction deviation of 1.254 to 4.933. These results clearly demonstrate that F_v /F_m , the widely used ChlF induction feature, can be determined from measurements without dark adaptation of samples. This will not only save experiment time but also make F_v /F_m useful in real-time and field applications. This work provides a high-throughput method to determine the important photosynthetic feature through ChlF for phenotyping plants.

Plant diseases threaten global food security by reducing crop yield; thus, diagnosing plant diseases is critical to agricultural production. Artificial intelligence technologies gradually replace traditional plant disease diagnosis methods due to their time-consuming, costly, inefficient, and subjective disadvantages. As a mainstream AI method, deep learning has substantially improved plant disease detection and diagnosis for precision agriculture. In the meantime, most of the existing plant disease diagnosis methods usually adopt a pre-trained deep learning model to support diagnosing diseased leaves. However, the commonly used pre-trained models are from the computer vision dataset, not the botany dataset, which barely provides the pre-trained models sufficient domain knowledge about plant disease. Furthermore, this pre-trained way makes the final diagnosis model more difficult to distinguish between different plant diseases and lowers the diagnostic precision. To address this issue, we propose a series of commonly used pre-trained models based on plant disease images to promote the performance of disease diagnosis. In addition, we have experimented with the plant disease pre-trained model on plant disease diagnosis tasks such as plant disease identification, plant disease detection, plant disease segmentation, and other subtasks. The extended experiments prove that the plant disease pre-trained model can achieve higher accuracy than the existing pre-trained model with less training time, thereby supporting the better diagnosis of plant diseases. In addition, our pre-trained models will be open-sourced at https://pd.samlab.cn/ and Zenodo platform https://doi.org/10.5281/zenodo.7856293.

The tea yield estimation provides information support for the harvest time and amount and serves as a decision-making basis for farmer management and picking. However, the manual counting of tea buds is troublesome and inefficient. To improve the efficiency of tea yield estimation, this study presents a deep-learning-based approach for efficiently estimating tea yield by counting tea buds in the field using an enhanced YOLOv5 model with the Squeeze and Excitation Network. This method combines the Hungarian matching and Kalman filtering algorithms to achieve accurate and reliable tea bud counting. The effectiveness of the proposed model was demonstrated by its mean average precision of 91.88% on the test dataset, indicating that it is highly accurate at detecting tea buds. The model application to the tea bud counting trials reveals that the counting results from test videos are highly correlated with the manual counting results (R ² = 0.98), indicating that the counting method has high accuracy and effectiveness. In conclusion, the proposed method can realize tea bud detection and counting in natural light and provides data and technical support for rapid tea bud acquisition.