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.

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.

Accurate detection of emerged seedlings is essential for intelligent management in the context of industrialized pepper seeding cultivation. Manual assessment of seedling emergence rates is inefficient and subjective. Meanwhile, dense seedling distributions and complex substrate backgrounds in plug trays often cause false positives and missed detections in vision-based systems. To address these issues, Adaptive-Dynamic-Spatial YOLO (ADS-YOLO), a compact detection model, was built based upon YOLO11n. First, an Adaptive Scale Encoding (ASE) module was introduced into the neck to preserve fine-grained spatial information through parallel variable-scale processing and fusion. Second, a Dynamic Scale Sequence Fusion (DYSF) module combined content-aware DySample upsampling with a scale-indexed three-dimensional feature representation to improve multiscale alignment and fusion. Third, a Spatial-Channel Position Attention (SCPA) module was used to emphasize seedling-related responses and suppress interference from substrate particles, water stains, and reflections. On the pepper seedling validation set obtained from a fixed 8:1:1 image-level data split, ADS-YOLO achieved a precision of 79.0%, a recall of 75.5%, and an mAP@0.5 of 78.1%, representing improvements of 3.3, 3.2, and 2.9 percentage points over YOLO11n. The model contained approximately 3.04 M parameters and required 8.1 GFLOPs. Furthermore, a prototype based on offline video-frame analysis was developed to display seedling detections and frame-level counts. The prototype achieved an offline processing throughput of 24.5 FPS and a per-frame end-to-end processing time of 40.8 ms. These measurement results indicate that the method proposed in this paper can be used in practical scenarios, and also lay the foundation for future real-time tracking and counting.