-
灰葡萄孢菌Botrytis cinerea是引起番茄灰霉病的主要病原真菌[1]，造成采后番茄Solanum lycopersicum果实软腐[2]、霉层形成[3]及风味劣变[4]。侵染早期可见症状不明显[5]，而色素、水分、细胞壁及代谢已发生变化[6−7]，使人工巡检难以及时识别，需发展快速、无损的早期检测方法[8−10]。
近年来，围绕果蔬采后品质与病害无损检测，已有若干相关综述发表：JELALI[11]和DAS等[12]分别综述了番茄病害深度学习检测，但主要聚焦于叶片尺度的田间场景。ALI等[13]综述了电子鼻在采后果蔬病害早期检测中的应用，但未涉及与其他传感模态的融合。OLADIPO等[14]综述了番茄采后保鲜技术，但未系统讨论无损检测技术体系。与上述工作不同，本研究以采后番茄灰霉病为主线，系统整合病理机制与多模态无损检测研究，重点探讨灰霉病早期可感知信号的生物学基础，不同检测模态的适用边界，及每种独立模态在多模态融合框架中的协同作用，并明确识别当前研究从叶片向果实、从实验室向产线迁移中的关键瓶颈与未来方向。
-
番茄灰霉病具有明显的潜伏侵染特征，病原可在田间初始侵染后以静止状态存在，也可在采后经机械损伤或果面裂口再次侵入[15]。当果实进入后熟、衰老或高湿储运环境时，潜伏病菌会迅速转入活动病斑阶段并持续扩展[16]。从病理过程看，病原早期侵染依赖伤口、角质层薄弱区及细胞壁降解过程，这也是其由潜伏态转入活动病斑阶段的重要基础。ROMANAZZI等[17]进一步指出，灰葡萄孢菌在低温储藏条件下仍可维持缓慢的菌丝扩展和组织侵染活性，低温储藏仅能延缓病程进展而无法彻底阻断侵染链条。因此，番茄灰霉病的发生实质上是从田间到储运期间的连续过程，这也是早期难以凭外观识别的主要原因[18]。
-
灰葡萄孢菌能够主动促进番茄后熟进程，从而增强致病能力。SILVA等[7]通过信使核糖核酸(mRNA)测序证实，灰葡萄孢菌侵染可上调番茄乙烯生物合成基因表达，加速果实转色和质地软化；基于酶联免疫吸附测定(ELISA)的糖组学技术揭示，侵染果实的细胞壁多糖变化与自然成熟果实高度相似。在未熟果实接种实验中，果胶酶基因(Bcpg1/Bcpg2)的双基因敲除突变体后，即使果实随后成熟，也仍不能由潜伏态转为活跃侵染，说明果胶降解是灰葡萄孢菌启动侵染和诱导宿主成熟的必要条件。因此，“诱导宿主成熟—促进细胞壁解聚—增强病原扩展”构成了采后灰霉病发展的正反馈机制。
-
在采后番茄灰霉病的病程发展过程中，番茄果实已出现色泽、硬度、呼吸代谢和挥发性代谢等多维变化。这些变化构成了不同检测模态的生物学靶点：色素衰减与组织结构变化为光谱检测提供了反射或吸收特征响应基础；表面褪色、水渍化为机器视觉检测提供了可见表面特征；挥发性有机化合物组成与丰度的代谢重编程为电子鼻检测提供了气味指纹信号来源。上述多维信号的异步性和互补性为不同模态间互补提供依据。
-
正常成熟与灰葡萄孢菌侵染均会改变果实色素、组织结构及含水状态，光谱判别需结合正常成熟基线、局部空间异常和多波段特征。MA等[19]利用时间分辨叶绿素荧光寿命，监测果实成熟与早期衰败，为时序基线建模提供参照。JAVIDAN等[20]结合图像与高光谱识别番茄叶片真菌病害；KAPETAS等[21]结合五波段图像、自注意力模型和实时荧光定量聚合酶链式反应，实现番茄叶片灰霉病早期检测。
现有研究仍以番茄叶片为主，采后果实应用需进一步控制成熟度差异。对草莓Fragaria × ananassa叶片和果实的研究也支持光谱识别早期真菌侵染[22−26]，但相应波段和判别阈值仍需在番茄果实上验证。
-
XIE等[8]利用高光谱区分健康番茄叶片与不同感染程度，FAHRENTRAPP等[9]采用五波段传感器监测番茄叶片，在受控条件下最早于接种后9 h识别感染区域，720 nm红边和840 nm近红外通道判别力较高。这些结果提供了病程分级及少波段早期检测的依据，但仍需在采后果实中验证。
番茄品质研究已覆盖成熟度、硬度、糖酸和色素预测[27−30]，并延伸至采前监测、便携式和手持检测[31−34]。低温胁迫番茄幼苗研究展示了由高光谱筛选波段、再简化为多光谱的路线[35]；苹果Malus domestica霉心病[36]和表面缺陷[37]、马铃薯Solanum tuberosum黑心病[38]的光谱检测也提供了方法参照。代表性工作见表1。
表 1 光谱技术在番茄相关研究中的代表性工作
Table 1. Representative works on the application of spectroscopic techniques in the study of tomato
-
机器视觉通过颜色、形状和纹理进行番茄成熟度、缺陷及病害识别，成本低，易于在线部署，但对采后灰霉病症前检测研究仍有限[41−42]。传统阈值与几何方法易受光照、遮挡和个体差异影响；单阶段目标检测模型(YOLO)及轻量化网络改善了成熟度与外观检测，并向实时部署发展[43−46]。
现有高精度结果多来自受控场景，储运中的反光、冷凝水和遮挡仍影响泛化[11−12,47]。常规彩色视觉适合症状出现后的在线筛查，潜伏侵染的内部生理变化则需结合光谱等信息。代表性工作见表2。
表 2 机器视觉在番茄灰霉病及相关任务中的代表性研究
Table 2. Representative studies of machine vision in tomato gray mold and related tasks research
研究对象和任务 方法类型 主要结果或结论 参考文献 番茄成熟度及外观品质识别 YOLOv5s、改进 YOLOv4/YOLOv5s-lite/
YOLOv7，兼顾识别、轻量化与实时性相较传统方法，深度学习在环境适应性和检测 精度上更具优势 [43−44] [45−46] 番茄外观缺陷检测 深度学习：残差网络(ResNet)迁移学习 可较有效地区分健康果与缺陷果，减少人工特 征工程依赖 [47] 番茄植株病害诊断 深度学习：显式控制隐藏类别策略 强调跨场景泛化问题，提示高精度结果未必可 直接迁移到复杂实际环境 [48] 棚室番茄果实识别 深度学习：融合注意力机制的识别模型 提高了棚室环境下果实识别模型的可靠性 [49] 灰霉病早期症状多光谱分割 语义分割网络U-Net++ 与MobileViT-S
编码器接种后第2天可定位病斑，总体准确率90.1% [50] 番茄灰霉病图像数据集构建 温室实拍标注数据集 含82张灰霉病图像，支持分类与目标检测任务 [51] 此外，GIAKOUMOGLOU等[50]结合多光谱和深度学习分割灰霉病感染区域，在黄瓜Cucumis sativus植株接种后第2天获得0.375的交并比(intersection over union, IoU)，总体准确率为90.1%，为病程定位提供了方法参照。LIU等[51]发布包含番茄灰霉病的温室病害标注图像集，为模型训练提供了数据基础。
-
番茄在采后成熟、衰老及病原侵染过程中会持续发生挥发性有机化合物组成及丰度变化，使气味指纹成为连接品质响应与病害发生的前置信号。与侧重单体化合物解析的气相色谱-质谱联用技术(GC-MS)不同，电子鼻更擅长对整体气味模式进行无损、快速和低成本识别。
-
电子鼻最早主要用于表征番茄成熟相关的挥发性有机化合物变化。相关研究表明：电子鼻能够较清晰地区分不同成熟阶段的番茄样品[52]。这一结果提示了成熟过程会显著改变气味形式，因此成熟度是番茄病害电子鼻识别中必须控制的重要混杂因素。若不区分正常成熟程度导致的挥发性有机化合物漂移，就难以准确区分病原侵染引起的异常气味变化与自然后熟引起的正常气味变化。
电子鼻研究进一步由成熟度分级拓展到采后品质预测。已有研究利用电子鼻对硬度、可溶性固形物含量和pH等品质指标进行无损建模[53]。这说明挥发性特征还可用于果实品质指标的间接预测。
-
已有研究将电子鼻与机器视觉结合用于番茄成熟等级识别和硬度预测，并证明融合系统优于单一检测技术[54]。这说明气味信息与颜色、外观等视觉特征具有明显互补性。因为病害早期往往首先表现为代谢异常和气味偏移，而并非清晰可见的霉层或组织腐烂，因此电子鼻更适合承担前端预警角色，而视觉信息更适合后期症状确认与分级。
-
近年的工作开始直接面向与番茄灰霉病相关的挥发性有机物及其便携式检测。代表性研究开发了面向番茄成熟度、健康状态及灰葡萄孢菌感染状态识别的便携式多传感器系统，并证明其能够区分不同成熟度与感染状态[55]。更直接的证据来自针对番茄灰霉病抑制与气味监测的研究。该类研究揭示了电子鼻不仅能够区分与感染相关的异常气味，还可跟踪灰霉病导致的霉味缺陷与风味下降过程[56]。可见，电子鼻能够进一步刻画灰霉病造成的品质下降程度，是连接病害进程、风味响应和货架期变化的检测手段。
目前，该方向仍主要依赖理想实验条件和人工处理试验。抑菌处理研究显示，喜树碱对灰葡萄孢菌的菌丝生长及番茄离体叶片接种病斑具有抑制作用[1]。现有研究对自然潜伏侵染样本的识别能力及模型跨设备稳定性的验证仍不足。其他果蔬病害中的电子鼻与数据融合研究则为番茄灰霉病模型设计和后续多模态融合提供了方法参考。相关研究的代表性工作见表3。
表 3 番茄电子鼻检测的代表性研究
Table 3. Representative studies of electronic nose applications in tomatoes
电子鼻也已被用于马铃薯真菌性腐烂早检[57]、苹果霉心病识别[58]和梨腐败分级[59]。近年来关于番茄灰霉病相关挥发性有机化合物的生物学研究也为电子鼻检测提供了重要的机制支撑。SHI等[60]研究发现：出芽短梗霉菌Aureobasidium pullulans S2在抑制番茄灰霉病过程中可产生2-苯乙醇、2-庚醇等具有抑菌活性的挥发性有机化合物，说明灰霉病发生与防御过程中的挥发性有机化合物组分变化具有明确的生物学基础。ZHANG等[61]进一步证实，伯克霍尔德氏菌Burkholderia sp. BV6产生的挥发性有机化合物(尤其是2-壬醇)对樱桃番茄灰霉病具有显著抑制效果，且挥发性有机化合物的抑菌活性具有广谱性。YUN等[62]则从代谢通路层面揭示了暹罗芽孢杆菌Bacillus siamensis产生的挥发性有机化合物可通过改变灰葡萄孢菌的代谢途径和致病性来抑制樱桃番茄采后灰霉病，这为理解电子鼻信号的生物学可解释性提供了参考。
上述研究表明：番茄灰霉病的发生、发展和拮抗过程均伴随着特征性挥发性有机化合物组分的变化。这些研究为挥发性有机化合物相关检测提供了机制线索，其诊断价值仍需在受侵染果实中验证。可见，电子鼻作为挥发性有机化合物模式识别工具，在灰霉病早期预警中的应用具有明确的生物学依据。
-
现有番茄灰霉病检测融合研究主要涉及叶片病害与果实品质2个层面[63]。多光谱荧光-反射融合用于叶片灰霉病检测，早期识别精度与病程分类精度分别为97.21%和97.56%[41]；多维数据融合亦可用于番茄病叶分级[64]。果实研究中，图像、可见光-近红外光谱与触觉融合的成熟度识别准确率为99.4%，内外成熟不一致样品为94.4%，总体准确率较最优单模态提高5.2个百分点[63]；视觉与电子鼻融合的硬度预测决定系数(R2)为0.94[54]。上述结果分别支持病害检测和品质评价，尚不能直接等同于采后果实灰霉病早期检测性能。
-
其他果实研究中，可见光与高光谱融合用于番木瓜Carica papaya成熟度估计[65]；光谱与电子鼻结合已用于猕猴桃Actinidia deliciosa后熟品质跟踪[66]及苹果霉心病检测[67]，为番茄多模态检测提供了方法参考。融合增益取决于模态互补性、数据对齐和策略选择[68]；向灰霉病迁移时，仍需区分正常成熟与侵染信号。
其他对象的研究包括柚Citrus maxima果实外观与内部品质关联分析[69]，相关综述分别涉及食品电子鼻与电子舌应用，如苹果[70]、荔枝Litchi chinensis[71]、鸡蛋[72]和水稻Oryza sativa籽粒的无损品质检测[73]，以及蔬菜种子病害识别[74]；果蔬高光谱检测综述[75]提供了技术背景。肉类腐败的电子鼻-多光谱融合检测[76]，玉米Zea mays种子活力的交叉注意力融合研究[77]和番茄早疫病叶片色素高光谱估测[78]，分别为融合结构设计与病害生化特征表征提供参考。
-
基于现有研究，采后番茄灰霉病病程预测的最有潜力融合路线是构建以表型特征+高光谱内部组织特征+电子鼻特征为核心的三模态框架。该框架可以在以下4个层面协同设计：①数据层。以果实个体为单位进行时序对齐，同一果实在相同时间节点同步采集机器视觉图像、光谱数据和电子鼻响应信号，确保按果实个体和采集时点对应各模态数据，并使图像与光谱采样区域一致。②模型层。根据模态特性选择融合策略。对于常规彩色机器视觉图像与光谱等同源光学信号可采用特征级融合，将卷积神经网络提取的图像特征与光谱特征在隐层空间拼接或交叉注意力对齐；对于电子鼻等异源信号，可采用决策级融合，各模态独立建模后通过加权投票集成输出最终判别结果；近期提出的跨模态[77]注意力机制也为异源多模态端到端融合提供了新思路。③训练层。针对采后灰霉病样本获取困难的现实，引入数据增强、迁移学习和域自适应方法缓解小样本与跨设备漂移问题[79]。④应用层。以覆盖不同果型与品种、跨成熟度(绿熟至红熟)和跨储藏条件(常温与冷链)的稳健泛化为目标，逐步从实验室离线分析向冷链节点和分选线在线部署推进[80−81]。通过上述4个层面协同推进，就可能真正实现从“实验室验证”到“产业化应用”的跨越。
-
采后番茄灰霉病早期识别需要综合利用病原侵染、果实成熟和品质劣变产生的信号。
灰霉病的果实潜伏期侵染、成熟诱导易感和异常促熟共同决定了可见症状出现之前就已存在可利用的生物学信号，包括色素衰减、组织结构重塑和挥发性有机化合物代谢重编程。这是开展症前预警的理论基础。光谱、机器视觉和电子鼻分别对应内部组织与生化变化、表面表型信息以及挥发性代谢异常，但单一模态均难以独立覆盖灰霉病早期“外观滞后、内部先变、气味先漂移”的全过程。本综述的核心发现之一是当前研究在“采后果实尺度”和“灰霉病”2个关键词的交叉领域仍存在显著空白，多数光谱和视觉研究来源于叶片尺度或成熟度等邻近任务，直接面向番茄果实灰霉病早期检测的核心文献极为有限。多模态融合的真正价值在于重构从病理响应到检测模态识别之间的对应关系。已有证据表明，多光谱荧光-反射融合可实现灰霉病接种后早期分类，图像、光谱与触觉融合可改善番茄成熟度识别，电子鼻与视觉融合可用于番茄果实成熟等级识别和硬度预测。但当前融合研究普遍缺乏标准化的番茄果实时序数据集和跨场景验证。
未来研究应考虑推进以下方向：①数据集建设。构建标准化的番茄果实灰霉病时序数据集，涵盖不同品种、不同果实成熟阶段(绿熟期、转色期、粉红期、红熟期)和不同储藏条件(常温与冷链)，纳入接种后不同时间节点(如0、12、24、48、72 h)的多模态同步采集数据。②模型开发。发展病理机制约束的特征工程与可解释深度学习模型，使模型输出能够与灰霉病的生物学进程建立对应关系。③泛化能力提升。借助迁移学习、域自适应和数据增强技术，强化小样本、跨设备和跨品种条件下的模型泛化能力，缩小实验室与产线之间的性能差距。④工程化部署。推动便携式多传感融合装备开发，面向冷链节点和分选线场景，实现从离线分析向在线预警的转化。
本文已在中国知网网络首发，可在知网搜索、下载并阅读全文。
采后番茄灰霉病早期检测与品质评价的多模态传感研究进展
DOI: 10.11833/j.issn.2095-0756.20260234
Research progress on multimodal sensing for early detection of gray mold and quality assessment in postharvest tomatoes
-
摘要: 番茄灰霉病主要由灰葡萄孢菌Botrytis cinerea侵染引起。现有研究多侧重于番茄Solanum lycopersicum叶片尺度灰葡萄孢菌的田间检测或采用单一传感模态，尚缺乏以采后果实为核心，系统整合多检测模态并衔接病理机制的文献梳理。多模态传感技术可从多维信号捕捉病害早期微弱变化，已成为重要的早期预警手段。以“采后番茄灰霉” “光谱”“机器视觉”“电子鼻”“多模态融合”等为关键词组合，以近10 a研究为主，围绕病理机制、光谱、机器视觉、电子鼻及多模态融合梳理研究进展。成熟诱导易感，细胞壁降解及挥发性有机化合物变化为早期检测提供生物学基础。光谱可表征内部组织与生化变化，机器视觉适于表面筛查，电子鼻可感知气味变化。番茄成熟度评价中的多模态融合优势为灰霉病检测提供了方法参考。未来需构建覆盖不同果型、品种、成熟度与储藏条件的果实时序数据集，发展可解释融合模型并验证跨设备泛化能力，推进便携式装备与在线预警应用。表3参81Abstract: Existing studies primarily focus on field detection of gray mold caused by Botrytis cinerea at the leaf scale in tomatoes (Solanum lycopersicum) or employ a single sensing modality, lacking a systematic review that centers on postharvest fruit integrating multiple sensing modalities and linking them to pathological mechanisms. Multimodal sensing technology can capture subtle changes in early-stage disease from multidimensional signals and is an important means for early warning. Using keywords including “postharvest tomato gray mold”, “spectral imaging”, “machine vision”, “electronic nose” and “multimodal fusion” as combinations, this review summarizes research progress in pathological mechanisms, spectroscopy, machine vision, electronic noses, and multimodal fusion in the past 10 years. Mature-induced susceptibility, cell wall degradation, and changes in volatile organic compounds provide the biological basis for early detection. Spectroscopy characterizes internal tissues and biochemical changes. Machine vision supports surface screening, and electronic noses detect odor changes. The multimodal fusion benefits in tomato maturity assessment provide a methodological reference for gray mold detection. Future research should establish a real-time data set covering different fruit types, varieties, maturity stages, and storage conditions, develop interpretable fusion models and verify cross-device generalization ability, and promote portable sensing equipment and online early-warning applications. [Ch, 3 tab. 81 ref.]
-
Key words:
- tomato (Solanum lycopersicum) /
- gray mold /
- spectral imaging /
- machine vision /
- electronic nose /
- multimodal fusion
-
表 1 光谱技术在番茄相关研究中的代表性工作
Table 1. Representative works on the application of spectroscopic techniques in the study of tomato
下载: 导出CSV
表 2 机器视觉在番茄灰霉病及相关任务中的代表性研究
Table 2. Representative studies of machine vision in tomato gray mold and related tasks research
研究对象和任务 方法类型 主要结果或结论 参考文献 番茄成熟度及外观品质识别 YOLOv5s、改进 YOLOv4/YOLOv5s-lite/
YOLOv7，兼顾识别、轻量化与实时性相较传统方法，深度学习在环境适应性和检测 精度上更具优势 [43−44] [45−46] 番茄外观缺陷检测 深度学习：残差网络(ResNet)迁移学习 可较有效地区分健康果与缺陷果，减少人工特 征工程依赖 [47] 番茄植株病害诊断 深度学习：显式控制隐藏类别策略 强调跨场景泛化问题，提示高精度结果未必可 直接迁移到复杂实际环境 [48] 棚室番茄果实识别 深度学习：融合注意力机制的识别模型 提高了棚室环境下果实识别模型的可靠性 [49] 灰霉病早期症状多光谱分割 语义分割网络U-Net++ 与MobileViT-S
编码器接种后第2天可定位病斑，总体准确率90.1% [50] 番茄灰霉病图像数据集构建 温室实拍标注数据集 含82张灰霉病图像，支持分类与目标检测任务 [51]
下载: 导出CSV
表 3 番茄电子鼻检测的代表性研究
Table 3. Representative studies of electronic nose applications in tomatoes
下载: 导出CSV
-
[1] 毛胜凤, 周湘, 范闻春, 等. 喜树碱对番茄灰霉病菌的毒理效应[J]. 浙江农林大学学报, 2015, 32(4): 585−589. MAO Shengfeng, ZHOU Xiang, FAN Wenchun, et al. Toxicological effects of camptothecin to Botrytis cinerea[J]. Journal of Zhejiang A&F University, 2015, 32(4): 585−589. DOI: 10.11833/j.issn.2095-0756.2015.04.014. [2] WANG Zhenshuo, WU Mumian, LIAO Qinhong, et al. Current status and future trends of eco-friendly management of postharvest fungal decays in tomato fruit[J]. NPJ Science of Food, 2025, 9: 104. DOI: 10.1038/s41538-025-00477-w. [3] DEAN R, van KAN J A L, PRETORIUS Z A, et al. The top 10 fungal pathogens in molecular plant pathology[J]. Molecular Plant Pathology, 2012, 13(4): 414−430. DOI: 10.1111/j.1364-3703.2011.00783.x. [4] CANTU D, BLANCO-ULATE B, YANG Liya, et al. Ripening-regulated susceptibility of tomato fruit to Botrytis cinerea requires NOR but not RIN or ethylene[J]. Plant Physiology, 2009, 150(3): 1434−1449. DOI: 10.1104/pp.109.138701. [5] SHAH P, POWELL A L T, ORLANDO R, et al. Proteomic analysis of ripening tomato fruit infected by Botrytis cinerea[J]. Journal of Proteome Research, 2012, 11(4): 2178−2192. DOI: 10.1021/pr200965c. [6] TZORTZAKIS N. Physiological and proteomic approaches to address the active role of Botrytis cinerea inoculation in tomato postharvest ripening[J]. Microorganisms, 2019, 7(12): 681. DOI: 10.3390/microorganisms7120681. [7] SILVA C J, ADASKAVEG J A, MESQUIDA-PESCI S D, et al. Botrytis cinerea infection accelerates ripening and cell wall disassembly to promote disease in tomato fruit[J]. Plant Physiology, 2023, 191(1): 575-590. DOI: 10.1093/plphys/kiac408. [8] XIE Chuanqi, YANG Ce, HE Yong. Hyperspectral imaging for classification of healthy and gray mold diseased tomato leaves with different infection severities[J]. Computers and Electronics in Agriculture, 2017, 135: 154−162. DOI: 10.1016/j.compag.2016.12.015. [9] FAHRENTRAPP J, RIA F, GEILHAUSEN M, et al. Detection of gray mold leaf infections prior to visual symptom appearance using a five-band multispectral sensor[J]. Frontiers in Plant Science, 2019, 10: 628. DOI: 10.3389/fpls.2019.00628. [10] HAGHBIN N, BAKHSHIPOUR A, ZAREIFOROUSH H, et al. Non-destructive pre-symptomatic detection of gray mold infection in kiwifruit using hyperspectral data and chemometrics[J]. Plant Methods, 2023, 19(1): 53. DOI: 10.1186/s13007-023-01032-y. [11] JELALI M. Deep learning networks-based tomato disease and pest detection: a first review of research studies using real field datasets[J]. Frontiers in Plant Science, 2024, 15: 1493322. DOI: 10.3389/fpls.2024.1493322. [12] DAS A, PATHAN F, JIM J R, et al. Deep learning-based classification, detection, and segmentation of tomato leaf diseases: a state-of-the-art review[J]. Artificial Intelligence in Agriculture, 2025, 15(2): 192−220. DOI: 10.1016/j.aiia.2025.02.006. [13] ALI A, MANSOL A S, KHAN A A, et al. Electronic nose as a tool for early detection of diseases and quality monitoring in fresh postharvest produce: a comprehensive review[J]. Comprehensive Reviews in Food Science and Food Safety, 2023, 22(3): 2408−2432. DOI: 10.1111/1541-4337.13151. [14] OLADIPO E K, OLADIPO B, OJUMU T V, et al. Postharvest losses of tomato: causes, spoilage mechanisms, and advances in conventional and emerging technologies for preservation[J]. Food Research International, 2026, 227: 118247. DOI: 10.1016/j.foodres.2025.118247. [15] HUANG Yuping, LI Ziang, BIAN Zhouchen, et al. Overview of deep learning and nondestructive detection technology for quality assessment of tomatoes[J]. Foods, 2025, 14(2): 286. DOI: 10.3390/foods14020286. [16] SPADA M, PUGLIESI C, FAMBRINI M, et al. Challenges and opportunities arising from host-Botrytis cinerea interactions to outline novel and sustainable control strategies: the key role of RNA interference[J]. International Journal of Molecular Sciences, 2024, 25(12): 6798. DOI: 10.3390/ijms25126798. [17] ROMANAZZI G, SMILANICK J L, FELIZIANI E, et al. Integrated management of postharvest gray mold on fruit crops[J]. Postharvest Biology and Technology, 2016, 113: 69−76. DOI: 10.1016/j.postharvbio.2015.11.003. [18] CHEN Jianhua, CAO Kexin, LU Xuan, et al. Investigating the action model of the resistance enhancement induced by bacterial volatile organic compounds against Botrytis cinerea in tomato fruit[J]. Frontiers in Plant Science, 2024, 15: 1475416. DOI: 10.3389/fpls.2024.1475416. [19] MA Te, INAGAKI T, TSUCHIKAWA S. Development of a sensitivity-enhanced chlorophyll fluorescence lifetime spectroscopic method for nondestructive monitoring of fruit ripening and postharvest decay[J]. Postharvest Biology and Technology, 2023, 198: 112231. DOI: 10.1016/j.postharvbio.2022.112231. [20] JAVIDAN S M, BANAKAR A, VAKILIAN K A, et al. Early detection and spectral signature identification of tomato fungal diseases (Alternaria alternata, Alternaria solani, Botrytis cinerea, and Fusarium oxysporum) by RGB and hyperspectral image analysis and machine learning[J]. Heliyon, 2024, 10(19): e38017. DOI: 10.1016/j.heliyon.2024.e38017. [21] KAPETAS D, KALOGEROPOULOU E, CHRISTAKAKIS P, et al. Multi-spectral image transformer descriptor classification combined with molecular tools for early detection of tomato grey mould[J]. Smart Agricultural Technology, 2024, 9: 100580. DOI: 10.1016/j.atech.2024.100580. [22] SIEDLISKA A, BARANOWSKI P, ZUBIK M, et al. Detection of fungal infections in strawberry fruit by VNIR/SWIR hyperspectral imaging[J]. Postharvest Biology and Technology, 2018, 139: 115−126. DOI: 10.1016/j.postharvbio.2018.01.018. [23] JIANG Qiyou, WU Gangshan, TIAN Chongfeng, et al. Hyperspectral imaging for early identification of strawberry leaves diseases with machine learning and spectral fingerprint features[J]. Infrared Physics & Technology, 2021, 118: 103898. DOI: 10.1016/j.infrared.2021.103898. [24] JUNG D H, DO KIM J, KIM H Y, et al. A hyperspectral data 3D convolutional neural network classification model for diagnosis of gray mold disease in strawberry leaves[J]. Frontiers in Plant Science, 2022, 13: 837020. DOI: 10.3389/fpls.2022.837020. [25] ZHANG Baohua, OU Yunmeng, YU Shuwan, et al. Gray mold and anthracnose disease detection on strawberry leaves using hyperspectral imaging[J]. Plant Methods, 2023, 19(1): 148. DOI: 10.1186/s13007-023-01123-w. [26] CHUN S W, SONG D J, LEE K H, et al. Deep learning algorithm development for early detection of Botrytis cinerea infected strawberry fruit using hyperspectral fluorescence imaging[J]. Postharvest Biology and Technology, 2024, 214: 112918. DOI: 10.1016/j.postharvbio.2024.112918. [27] ZHAO Mingrui, CANG Hao, CHEN Huixin, et al. Determination of quality and maturity of processing tomatoes using near-infrared hyperspectral imaging with interpretable machine learning methods[J]. LWT, 2023, 183: 114861. DOI: 10.1016/j.lwt.2023.114861. [28] SHAO Yuanyuan, SHI Yukang, QIN Yongdong, et al. A new quantitative index for the assessment of tomato quality using Vis-NIR hyperspectral imaging[J]. Food Chemistry, 2022, 386: 132864. DOI: 10.1016/j.foodchem.2022.132864. [29] DAI Chunxia, SUN Jun, HUANG Xingyi, et al. Application of hyperspectral imaging as a nondestructive technology for identifying tomato maturity and quantitatively predicting lycopene content[J]. Foods, 2023, 12(15): 2957. DOI: 10.3390/foods12152957. [30] XIANG Yun, CHEN Qijun, SU Zhongjing, et al. Deep learning and hyperspectral images based tomato soluble solids content and firmness estimation[J]. Frontiers in Plant Science, 2022, 13: 860656. DOI: 10.3389/fpls.2022.860656. [31] FASS E, SHLOMI E, ZIV C, et al. Machine learning models based on hyperspectral imaging for pre-harvest tomato fruit quality monitoring[J]. Computers and Electronics in Agriculture, 2025, 229: 109788. DOI: 10.1016/j.compag.2024.109788. [32] de BRITO A A, CAMPOS F, dos REIS NASCIMENTO A, et al. Non-destructive determination of color, titratable acidity, and dry matter in intact tomatoes using a portable Vis-NIR spectrometer[J]. Journal of Food Composition and Analysis, 2022, 107: 104288. DOI: 10.1016/j.jfca.2021.104288. [33] DUCKENA L, ALKSNIS R, ERDBERGA I, et al. Non-destructive quality evaluation of 80 tomato varieties using vis-NIR spectroscopy[J]. Foods, 2023, 12(10): 1990. DOI: 10.3390/foods12101990. [34] ALENAZI M M, SHAFIQ M, ALSADON A A, et al. Non-destructive assessment of flesh firmness and dietary antioxidants of greenhouse-grown tomato (Solanum lycopersicum L.) at different fruit maturity stages[J]. Saudi Journal of Biological Sciences, 2020, 27(10): 2839−2846. DOI: 10.1016/j.sjbs.2020.07.004. [35] KANG Y S, RYU C S, KANG J G. Presenting a multispectral image sensor for quantification of total polyphenols in low-temperature stressed tomato seedlings using hyperspectral imaging[J]. Sensors, 2024, 24(13): 4260. DOI: 10.3390/s24134260. [36] 王梓萌, 刘景艳, 姚腾飞, 等. 基于近红外光谱技术的苹果霉心病检测方法[J]. 科学技术与工程, 2019, 19(10): 150−154. WANG Zimeng, LIU Jingyan, YAO Tengfei, et al. Detection method of moldy core in apple fruit based on near infrared spectroscopy[J]. Science Technology and Engineering, 2019, 19(10): 150−154. DOI: 10.3969/j.issn.1671-1815.2019.10.023. [37] 孟庆龙, 张艳, 尚静. 基于高光谱成像技术的苹果表面缺陷无损检测[J]. 食品研究与开发, 2019, 40(5): 168−172. MENG Qinglong, ZHANG Yan, SHANG Jing. Nondestructive detection of defect on apples using hyperspectral imaging technology[J]. Food Research and Development, 2019, 40(5): 168−172. DOI: 10.3969/j.issn.1005-6521.2019.05.030. [38] 韩亚芬, 吕程序, 苑严伟, 等. PLS-DA优化模型的马铃薯黑心病可见近红外透射光谱检测[J]. 光谱学与光谱分析, 2021, 41(4): 1213−1219. HAN Yafen, LÜ Chengxu, YUAN Yanwei, et al. PLS-discriminant analysis on potato blackheart disease based on VIS-NIR transmission spectroscopy[J]. Spectroscopy and Spectral Analysis, 2021, 41(4): 1213−1219. DOI: 10.3964/j.issn.1000-0593(2021)04-1213-07. [39] KONG Wenwen, LIU Fei, ZHANG Chu, et al. Fast detection of peroxidase (POD) activity in tomato leaves which infected with Botrytis cinerea using hyperspectral imaging[J]. Spectrochimica Acta Part A, Molecular and Biomolecular Spectroscopy, 2014, 118: 498−502. DOI: 10.1016/j.saa.2013.09.009. [40] 袁娉, 冯洁, 张江坤, 等. 基于高光谱的番茄病毒病分级检测识别[J]. 光散射学报, 2025, 37(4): 813−821. YUAN Ping, FENG Jie, ZHANG Jiangkun, et al. Study on classification detection and recognition of tomato virnus diseases based on hyperspectral images[J]. The Journal of Light Scattering, 2025, 37(4): 813−821. DOI: 10.13883/j.issn1004-5929.202504036. [41] ZHONG Xiaohao, LI Huicheng, CAI Yixin, et al. Early detection of tomato gray mold based on multispectral imaging and machine learning[J]. Horticulturae, 2025, 11(9): 1073. DOI: 10.3390/horticulturae11091073. [42] 戴军. 机器视觉技术在瓜菜检测应用中的研究进展[J]. 中国瓜菜, 2023, 36(11): 1−9. DAI Jun. Research progress of machine vision technology in the detection of cucur-bits and vegetables[J]. China Cucurbits and Vegetables, 2023, 36(11): 1−9. DOI: 10.3969/j.issn.1673-2871.2023.11.001. [43] 陈伟, 张春雨, 朱超冉. 基于YOLOv5s算法的番茄成熟度识别研究[J]. 安徽科技学院学报, 2023, 37(1): 92−95. CHEN Wei, ZHANG Chunyu, ZHU Chaoran. Research on tomato maturity recognition based on YOLOv5s algorithm[J]. Journal of Anhui Science and Technology University, 2023, 37(1): 92−95. DOI: 10.19608/j.cnki.1673-8772.2023.0012. [44] 吕金锐, 付燕, 倪美玉, 等. 基于改进YOLOv4模型的番茄成熟度检测方法[J]. 食品与机械, 2023, 39(9): 134−139. LÜ Jinrui, FU Yan, NI Meiyu, et al. Research on tomato maturity detection method based on improved YOLOv4 model[J]. Food & Machinery, 2023, 39(9): 134−139. DOI: 10.13652/j.spjx.1003.5788.2023.60046. [45] 孙宇朝, 李守豪, 夏秀波, 等. 利用改进YOLOv5s模型检测番茄果实成熟度及外观品质[J]. 园艺学报, 2024, 51(2): 396−410. SUN Yuchao, LI Shouhao, XIA Xiubo, et al. Detecting tomato fruit ripeness and appearance quality based on improved YOLOv5s[J]. Acta Horticulturae Sinica, 2024, 51(2): 396−410. DOI: 10.16420/j.issn.0513-353x.2023-0859. [46] 苗荣慧, 李志伟, 武锦龙. 基于改进YOLO v7的轻量化樱桃番茄成熟度检测方法[J]. 农业机械学报, 2023, 54(10): 225−233. MIAO Ronghui, LI Zhiwei, WU Jinlong. Lightweight maturity detection of cherry tomato based on improved YOLO v7[J]. Transactions of the Chinese Society for Agricultural Machinery, 2023, 54(10): 225−233. DOI: 10.6041/j.issn.1000-1298.2023.10.022. [47] da COSTA A Z, FIGUEROA H E H, FRACAROLLI J A. Computer vision based detection of external defects on tomatoes using deep learning[J]. Biosystems Engineering, 2020, 190: 131−144. DOI: 10.1016/j.biosystemseng.2019.12.003. [48] FUENTES A, YOON S, LEE M H, et al. Improving accuracy of tomato plant disease diagnosis based on deep learning with explicit control of hidden classes[J]. Frontiers in Plant Science, 2021, 12: 682230. DOI: 10.3389/fpls.2021.682230. [49] 陈翠琴, 孟清. 基于图像处理技术的棚室番茄果实识别[J]. 农机化研究, 2025, 47(1): 189−193. CHEN Cuiqin, MENG Qing. Tomato fruit recognition in greenhouse based on image processing technology[J]. Journal of Agricultural Mechanization Research, 2025, 47(1): 189−193. DOI: 10.13427/j.issn.1003-188X.2025.01.030. [50] GIAKOUMOGLOU N, KALOGEROPOULOU E, KLARIDOPOULOS C, et al. Early detection of Botrytis cinerea symptoms using deep learning multi-spectral image segmentation[J]. Smart Agricultural Technology, 2024, 8: 100481. DOI: 10.1016/j.atech.2024.100481. [51] LIU Yongbo, ZHU Yuhang, HU Liang, et al. A labeled image dataset of common tomato diseases for classification and object detection[J]. Data in Brief, 2025, 63: 112032. DOI: 10.1016/j.dib.2025.112032. [52] GÓMEZ A H, HU Guixian, WANG Jun, et al. Evaluation of tomato maturity by electronic nose[J]. Computers and Electronics in Agriculture, 2006, 54(1): 44−52. DOI: 10.1016/j.compag.2006.07.002. [53] FENG Lei, ZHANG Min, BHANDARI B, et al. A novel method using MOS electronic nose and ELM for predicting postharvest quality of cherry tomato fruit treated with high pressure Argon[J]. Computers and Electronics in Agriculture, 2018, 154: 411−419. DOI: 10.1016/j.compag.2018.09.032. [54] HUANG Xingyi, PAN Sihui, SUN Zhaoyan, et al. Evaluating quality of tomato during storage using fusion information of computer vision and electronic nose[J]. Journal of Food Process Engineering, 2018, 41(6): e12832. DOI: 10.1111/jfpe.12832. [55] MELÉNDEZ F, SÁNCHEZ R, FERNÁNDEZ J Á, et al. Design of a multisensory device for tomato volatile compound detection based on a mixed metal oxide—electrochemical sensor array and optical reader[J]. Micromachines, 2023, 14(9): 1761. DOI: 10.3390/mi14091761. [56] BAREA-RAMOS J D, RODRÍGUEZ M J, CALVO P, et al. Inhibition of Botrytis cinerea in tomatoes by allyl-isothiocyanate release from black mustard (Brassica nigra) seeds and detection by E-nose[J]. Food Chemistry, 2024, 432: 137222. DOI: 10.1016/j.foodchem.2023.137222. [57] 黄星奕, 孙兆燕, 田潇瑜, 等. 基于电子鼻技术的马铃薯真菌性腐烂病早期检测[J]. 食品工业科技, 2018, 39(24): 97−101. HUANG Xingyi, SUN Zhaoyan, TIAN Xiaoyu, et al. Early detection of potato rot disease caused by fungal based on electronic nose technology[J]. Science and Technology of Food Industry, 2018, 39(24): 97−101. DOI: 10.13386/j.issn1002-0306.2018.24.018. [58] 张建超, 张鹏, 薛友林, 等. 基于电子鼻表征霉心病苹果特征气味及无损检测模型建立[J]. 食品与发酵工业, 2022, 48(2): 267−273. ZHANG Jianchao, ZHANG Peng, XUE Youlin, et al. Characterization of characteristic odor and establishment of nondestructive detection model of core rot apples based on electronic nose[J]. Food and Fermentation Industries, 2022, 48(2): 267−273. DOI: 10.13995/j.cnki.11-1802/ts.026717. [59] 赵策, 马飒飒, 张磊, 等. 基于电子鼻技术的皇冠梨腐败等级分类研究[J]. 食品工业科技, 2020, 41(3): 246−250, 258. ZHAO Ce, MA Sasa, ZHANG Lei, et al. Research on classification of rotten grades of Huangguan pears on electronic nose technology[J]. Science and Technology of Food Industry, 2020, 41(3): 246−250, 258. DOI: 10.13386/j.issn1002-0306.2020.03.041. [60] SHI Yu, ZHAO Qianhua, XIN Yu, et al. Aureobasidium pullulans S2 controls tomato gray mold and produces volatile organic compounds and biofilms[J]. Postharvest Biology and Technology, 2023, 204: 112450. DOI: 10.1016/j.postharvbio.2023.112450. [61] ZHANG Yu, CHEN Minhua, SANG Jierong, et al. Antifungal volatile organic compounds from Burkholderia sp. BV6 control gray mold disease of cherry tomato[J]. Biological Control, 2025, 211: 105935. DOI: 10.1016/j.biocontrol.2025.105935. [62] YUN Jing, YAN Xin, YUAN Shuzhi, et al. Volatile organic compounds produced by Bacillus siamensis inhibit postharvest grey mold in cherry tomato by altering metabolic pathways and the pathogenicity of Botrytis cinerea[J]. Postharvest Biology and Technology, 2025, 226: 113543. DOI: 10.1016/j.postharvbio.2025.113543. [63] LIU Yang, WEI Chaojie, YOON S C, et al. Development of multimodal fusion technology for tomato maturity assessment[J]. Sensors, 2024, 24(8): 2467. DOI: 10.3390/s24082467. [64] 袁娉, 冯洁, 王少波, 等. 多维数据融合检测识别番茄病害[J]. 光散射学报, 2025, 37(4): 822−831. YUAN Ping, FENG Jie, WANG Shaobo, et al. The detection and identification of tomato diseases through multi-dimensional data fusion[J]. The Journal of Light Scattering, 2025, 37(4): 822−831. DOI: 10.13883/j.issn1004-5929.202504037. [65] GARILLOS-MANLIGUEZ C A, CHIANG J Y. Multimodal deep learning and visible-light and hyperspectral imaging for fruit maturity estimation[J]. Sensors, 2021, 21(4): 1288. DOI: 10.3390/s21041288. [66] BAKHSHIPOUR A. A data fusion approach for nondestructive tracking of the ripening process and quality attributes of green Hayward kiwifruit using artificial olfaction and proximal hyperspectral imaging techniques[J]. Food Science & Nutrition, 2023, 11(10): 6116−6132. DOI: 10.1002/fsn3.3548. [67] 杨晨昱, 袁鸿飞, 马惠玲, 等. 基于傅里叶近红外光谱和电子鼻技术的苹果霉心病无损检测[J]. 食品与发酵工业, 2021, 47(7): 211−216. YANG Chenyu, YUAN Hongfei, MA Huiling, et al. Nondestructive detection of apple moldy core based on FT-NIR and electronic nose technology[J]. Food and Fermentation Industries, 2021, 47(7): 211−216. DOI: 10.13995/j.cnki.11-1802/ts.025671. [68] VANARAJ R, I P B, MAYAKRISHNAN G, et al. A systematic review of the applications of electronic nose and electronic tongue in food quality assessment and safety[J]. Chemosensors, 2025, 13(5): 161. DOI: 10.3390/chemosensors13050161. [69] 徐赛, 陆华忠, 王旭, 等. 柚果视觉特征与内部品质的相关性及粗略检测方法研究[J]. 食品与发酵工业, 2021, 47(24): 242−248. XU Sai, LU Huazhong, WANG Xu, et al. Research on relationship between vision features and internal quality parameters of pomelo and a rough detection method for pomelo[J]. Food and Fermentation Industries, 2021, 47(24): 242−248. DOI: 10.13995/j.cnki.11-1802/ts.027138. [70] 贾一鸣, 朱丽霞, 张锐利. 无损检测技术在苹果品质检测中的应用[J]. 食品科技, 2021, 46(4): 262−267. JIA Yiming, ZHU Lixia, ZHANG Ruili. Application of nondestructive testing technology in apple quality testing[J]. Food Science and Technology, 2021, 46(4): 262−267. DOI: 10.13684/j.cnki.spkj.2021.04.041. [71] 谢天华, 林静, 梁峻槐, 等. 无损检测技术在荔枝品质检测中的应用与发展[J]. 农技服务, 2025, 42(3): 70−75. XIE Tianhua, LIN Jing, LIANG Junhuai, et al. Application and development of nondestructive testing technology in Litchi quality inspection[J]. Agricultural Technology Service, 2025, 42(3): 70−75. [72] 王明, 于金莹, 胡雁翔, 等. 无损检测技术在鸡蛋品质检测中的应用研究进展[J]. 食品科技, 2021, 46(4): 268−272. WANG Ming, YU Jinying, HU Yanxiang, et al. Application of nondestructive testing technology in detection of egg quality[J]. Food Science and Technology, 2021, 46(4): 268−272. DOI: 10.13684/j.cnki.spkj.2021.04.042. [73] 孟宁, 刘明, 刘艳香, 等. 无损检测技术在稻米品质检测中的应用及展望[J]. 粮油食品科技, 2019, 27(6): 98−101. MENG Ning, LIU Ming, LIU Yanxiang, et al. Application and prospect of nondestructive testing technology in rice quality detection[J]. Science and Technology of Cereals, Oils and Foods, 2019, 27(6): 98−101. DOI: 10.16210/j.cnki.1007-7561.2019.06.018. [74] 陈广宇, 王关平, 李彤, 等. 蔬菜种子病害及其识别方法的研究进展[J]. 林业机械与木工设备, 2023, 51(11): 4−10. CHEN Guangyu, WANG Guanping, LI Tong, et al. Research progress on vegetable seed diseases and their identification methods[J]. Forestry Machinery & Woodworking Equipment, 2023, 51(11): 4−10. DOI: 10.13279/j.cnki.fmwe.2023.0077. [75] 卢娜, 韩平, 王纪华. 高光谱成像技术在果蔬品质安全无损检测中的应用[J]. 食品安全质量检测学报, 2017, 8(12): 4594−4601. LU Na, HAN Ping, WANG Jihua. Application of hyperspectral imaging in non-destructive testing of fruits and vegetables[J]. Journal of Food Safety & Quality, 2017, 8(12): 4594−4601. DOI: 10.3969/j.issn.2095-0381.2017.12.016. [76] KODOGIANNIS V S, ALSHEJARI A. Data fusion of electronic nose and multispectral imaging for meat spoilage detection using machine learning techniques[J]. Sensors, 2025, 25(10): 3198. DOI: 10.3390/s25103198. [77] XIE Jieming, ZHENG Jian, FU Jiacheng, et al. Research on multi modal corn seed vitality grading based on three branch cross attention fusion network[J]. Spectrochimica Acta Part A, Molecular and Biomolecular Spectroscopy, 2026, 358: 127828. DOI: 10.1016/j.saa.2026.127828. [78] 尹小君, 赵思峰. 加工番茄早疫病病叶色素含量高光谱估测[J]. 遥感信息, 2012, 27(3): 40−45. YIN Xiaojun, ZHAO Sifeng. Hyperspectral estimation of pigment content of process tomato leaves under early blight[J]. Remote Sensing Information, 2012, 27(3): 40−45. DOI: 10.3969/j.issn.1000-3177.2012.03.007. [79] CHEN Lei, ZOU Jiaxian, YUAN Yuan, et al. Improved domain adaptive rice disease image recognition based on a novel attention mechanism[J]. Computers and Electronics in Agriculture, 2023, 208: 107806. DOI: 10.1016/j.compag.2023.107806. [80] GIRMATSION M, TANG Xiaoqian, ZHANG Qi, et al. Progress in machine learning-supported electronic nose and hyperspectral imaging technologies for food safety assessment: a review[J]. Food Research International, 2025, 209: 116285. DOI: 10.1016/j.foodres.2025.116285. [81] SINGH R, NISHA R, NAIK R, et al. Sensor fusion techniques in deep learning for multimodal fruit and vegetable quality assessment: a comprehensive review[J]. Journal of Food Measurement and Characterization, 2024, 18(9): 8088−8109. DOI: 10.1007/s11694-024-02789-z. -
-
链接本文:
https://zlxb.zafu.edu.cn/article/doi/10.11833/j.issn.2095-0756.20260234
