Shi Xingren

目的夜间甜瓜识别和采摘严重依赖人工光源。不同人工补光配置下，甜瓜果实和茎部的图像表观可能不同，语义分割模型的训练和跨配置测试结果也可能存在差异。为助力夜间农产品识别和采摘，构建了多颜色人工光照甜瓜分割图像数据集MelonNightColor。方法在保持相机位置、拍摄距离和补光灯方位一致的条件下，切换补光灯红、绿、蓝通道配置，获取红、蓝、绿、青、紫、黄、白7种人工补光配置下的夜间甜瓜图像。当前已发布5474张图像及像素级标注（覆盖782个拍摄位置，标注背景、甜瓜果实和茎部3个类别）。以拍摄位置为最小划分单元划分训练集、验证集和测试集，确保同一场景的7种颜色图像不跨集合，并组织为7个单颜色子集和4个组合颜色子集。采用9种代表性语义分割模型进行基线评测，并以UNet构建7×7跨颜色泛化矩阵。结果9种模型在7种光源下的背景、果实和茎部的类别交并比（intersection over union，IoU）均值分别为99.05%、86.39%和38.84%，茎部因形态细长且易受遮挡影响而分割难度最大。白光和黄光下9种模型平均交并比（mean intersection over union，mIoU）均值较高，分别达78.17%和77.94%；红光和蓝光下相对较低，分别为72.93%和73.23%。UNet同颜色测试平均mIoU为80.38%，跨颜色测试平均mIoU为44.25%。该结果显示，在本数据集的设备、品种和采集条件下，不同补光配置及自动成像过程共同作用下，训练与测试补光配置不一致时模型性能出现明显下降。结论MelonNightColor提供了夜间甜瓜多颜色成组图像和像素级标注。未来，数据集MelonNightColor可为特定夜间人工补光条件下的甜瓜语义分割基线评测和跨配置模型研究提供公开数据基础。数据集已在中国科学院科学数据银行公开，数字对象标识符（digital object identifier，DOI）为https：//doi.org/10.57760/sciencedb.j00240.00237。

The accurate identification of tomato maturity and picking positions is essential for efficient picking. Current deep-learning models face challenges such as large parameter sizes, single-task limitations, and insufficient precision. This study proposes MTS-YOLO, a lightweight and efficient model for detecting tomato fruit bunch maturity and stem picking positions. We reconstruct the YOLOv8 neck network and propose the high- and low-level interactive screening path aggregation network (HLIS-PAN), which achieves excellent multi-scale feature extraction through the alternating screening and fusion of high- and low-level information while reducing the number of parameters. Furthermore, We utilize DySample for efficient upsampling, bypassing complex kernel computations with point sampling. Moreover, context anchor attention (CAA) is introduced to enhance the model’s ability to recognize elongated targets such as tomato fruit bunches and stems. Experimental results indicate that MTS-YOLO achieves an F1-score of 88.7% and an mAP@0.5 of 92.0%. Compared to mainstream models, MTS-YOLO not only enhances accuracy but also optimizes the model size, effectively reducing computational costs and inference time. The model precisely identifies the foreground targets that need to be harvested while ignoring background objects, contributing to improved picking efficiency. This study provides a lightweight and efficient technical solution for intelligent agricultural picking.