Estimating the mass of Oudemansiella raphanipies quickly and accurately is indispensable in optimizing post-harvest packaging processes. Traditional methods typically involve manual grading followed by weighing with a balance, which is inefficient and labor-intensive. To address the challenges encountered in actual production scenarios, in this work, we developed a novel pipeline for estimating the mass of multiple Oudemansiella raphanipies. To achieve this goal, an enhanced deep learning (DL) algorithm for instance segmentation and a machine learning (ML) model for mass prediction were introduced. On one hand, to segment multiple samples in the same image, a novel instance segmentation network named FinePoint-ORSeg was applied to obtain the finer edges of samples, by integrating an edge attention module to improve the fineness of the edges. On the other hand, for individual samples, a novel cap–stem segmentation approach was applied and 18 phenotypic parameters were obtained. Furthermore, principal component analysis (PCA) was utilized to reduce the redundancy among features. Combining the two aspects mentioned above, the mass was computed by an exponential GPR model with seven principal components. In terms of segmentation performance, our model outperforms the original Mask R-CNN; the $\mathrm{A}\mathrm{P}$, ${\mathrm{A}\mathrm{P}}_{50}$, ${\mathrm{A}\mathrm{P}}_{75}$, and ${\mathrm{A}\mathrm{P}}_{\mathrm{s}}$ are improved by 2%, 0.7%, 1.9%, and 0.3%, respectively. Additionally, our model outperforms other networks such as YOLACT, SOLOV2, and Mask R-CNN with Swin. As for mass estimation, the results show that the average coefficient of variation (CV) of a single sample mass in different attitudes is 6.81%. Moreover, the average mean absolute percentage error (MAPE) for multiple samples is 8.53%. Overall, the experimental results indicate that the proposed method is time-saving, non-destructive, and accurate. This can provide a reference for research on post-harvest packaging technology for Oudemansiella raphanipies.

Accurate and non-destructive volume estimation of agricultural products is essential for precision agriculture, yet remains challenging when transitioning from controlled laboratory conditions to complex orchard environments. Although 2D image-based volume estimation methods provide a cost-effective and scalable solution, existing studies are fragmented and lack a unified perspective on their real-world applicability. This review presents a systematic synthesis of 2D image-based volume estimation methods, explicitly framed through the laboratory-to-orchard transition. We categorized existing volume estimation approaches according to the sensing modality into monocular RGB-based approaches and depth-assisted methods, and further reviewed them based on the image processing methods. A key finding is that high-precision geometric estimation can be achieved in laboratory environments, whereas deep learning and RGB-D fusion have driven a shift from conventional geometric modeling toward data-driven and hybrid learning frameworks in orchard settings. However, 2D image-based volume estimation remains fundamentally limited by scale ambiguity, severe occlusion, and sensitivity to illumination and background variability in real orchard environment. Overall, this review provides a unified perspective for understanding volume estimation methodology across environments and offers guidance for developing robust, scalable, and field-deployable volume estimation systems for real-world agricultural applications.

Estimating the mass of Oudemansiella raphanipies quickly and accurately is indispensable for optimizing post-harvest packaging processes. The traditional methods typically involve manual grading followed by weighing with a balance, which is inefficient and labor-intensive. To address the challenges encountered in actual production scenarios, in this work, we proposed a novel pipeline for estimating the mass of multiple Oudemansiella raphanipies. To achieve this goal, an enhanced deep learning (DL) algorithm for instance segmentation and a machine learning (ML) model for mass prediction were introduced. On one hand, to segment the multiple samples in the same image, a novel instance segmentation network named FinePoint-ORSeg was presented to obtain the finer edges of samples, which integrated the edge attention module for improving the fineness of the edges. On the other hand, for individual samples, a novel cap-stem segmentation approach was applied and 18 phenotypic parameters were obtained. Furthermore, the Principal Component Analysis (PCA) was utilized to reduce the redundancy among features. Combining the two aspects mentioned above, the mass was computed by Exponential GPR model with 7 principal components. In terms of segmentation performance, our model outperforms the original Mask R-CNN, the AP, the AP50, the AP75 and the APs are improved by 2%, 0.7%, 1.9%, and 0.3%, respectively. Additionally, our model outperforms other networks such as YOLACT, SOLOV2 and Mask R-CNN with swin. As for mass estimation, the results showed the average Coefficient of Variation (CV) of single sample mass in different attitude are 6.81%. Moreover, an average mean absolute percentage error (MAPE) of multiple samples is 8.53%. Overall, the experimental results indicated that the proposed method is time-saving, non-destructive and accurate. This can provide a reference for the research on post-harvest packaging technology of Oudemansiella raphanipies.

Monocular RGB imaging offers a low-cost, flexible approach to morphological measurement, quality assessment, growth monitoring, and production automation for edible mushroom fruiting bodies. However, the relationships among visible phenotypes, visual methods, measurement reliability, and production requirements remain insufficiently integrated. This problem-oriented review synthesizes 76 core studies published between 2016 and the final search date in 2026 through a framework linking visible phenotypes, visual tasks, key bottlenecks, and production applications. It covers individual localization and separation, structural measurement, spatial estimation, quality and species recognition, temporal analysis, and production deployment. The reviewed studies reveal a transition from static two-dimensional detection and counting toward instance-level morphological measurement, three-dimensional parameter estimation, and spatiotemporal growth modeling, extending phenotyping from visible appearance description to spatial trait estimation and growth prediction. Quantitative results also highlight the importance of evaluation conditions: for example, MSH-YOLOv8 achieved an ${\mathrm{AP}}_{50}$ of 98.49% on the Fungi dataset, whereas ${\mathrm{AP}}_{50:95}$ and small-object AP were 75.29% and 39.73%, respectively, indicating that high ${\mathrm{AP}}_{50}$ alone does not adequately characterize detection performance under stricter localization criteria or for small targets. These study-specific results cannot be directly extrapolated to commercial production environments. Severe occlusion, projection errors, inconsistent phenotype definitions, and temporal instability continue to constrain measurement reliability. Moreover, reliable performance without extensive retraining following changes in strains, substrates, or lighting systems remains insufficiently demonstrated. Future research should prioritize standardized multi-task and temporal datasets, unified phenotype definitions, uncertainty evaluation, occlusion-robust spatial and temporal modeling, and closed-loop production validation. Commercial scaling of low-cost monocular RGB imaging in protected mushroom cultivation depends on translating its affordability into reliable performance under severe occlusion and across production conditions, thereby enabling accessible automation for small and medium-scale producers.

Accurate detection of Camellia oleifera fruits in natural orchards is essential for automated harvesting and yield-related phenotyping but is challenged by small fruit size, dense distribution, foliage occlusion, and fruit–leaf similarity. To address these challenges, we developed CO-YOLO, an enhanced YOLOv11-based detector that integrates a spatial attention module (SAM) and coordinate attention (CA) to jointly enhance spatial discriminative feature extraction and positional encoding. Unlike approaches that employ attention mechanisms primarily to reweight features, CO-YOLO combines spatial attention with coordinate-aware feature representation to improve the identification and localization of partially occluded fruits while retaining the YOLOv11 detection framework. A field dataset of 674 raw RGB images was collected under different fruit developmental stages, weather conditions, illumination environments, and occlusion levels. Compared with YOLOv11, CO-YOLO improved precision, recall, mAP@0.50, mAP@0.75, mAP@0.50:0.95, and F1-score by 2.46, 3.17, 2.96, 3.64, 3.42, and 2.56 percentage points, respectively, with greater improvements under stringent localization criteria. Ablation experiments confirmed the complementary contributions of SAM and CA, while environmental evaluation showed that immature fruits, rainy conditions, and dense occlusion remained the main sources of performance degradation. Fruit-counting analysis further indicated improved consistency between automated and manual counts, although dense fruit clusters continued to introduce counting errors. Overall, CO-YOLO improves the detection of Camellia oleifera fruit under complex orchard conditions while preserving the efficiency of the YOLOv11 framework, and thus offers a practical visual perception approach for automated harvesting and yield-related phenotyping.