[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"item-2339":3},{"id":4,"title":5,"url":6,"summary":7,"summary_zh":8,"content":9,"source_name":10,"source_url":6,"published_at":11,"category":12,"cover_url":9,"hotness":13,"is_selected":14,"score":15,"score_detail":16,"sources":24,"tags":26,"view_count":32,"doi":33,"paper":34,"created_at":50},2339,"Quality Assessment of Scotch Bonnet Pepper (Capsicum Chinense Jacq) Using Image Processing Techniques","https:\u002F\u002Fdoi.org\u002F10.53982\u002Fajerd.2026.0903.06-j","Scotch Bonnet peppers (Capsicum chinense) are globally cultivated cash crops prized for their distinct culinary uses and high nutritional profile. However, meeting rising consumer demands for premium quality is hindered by traditional manual inspection, which remains subjective, slow, and labour-intensive. To address this, this study presents a computer vision-based framework designed to automate the pepper grading process. A convolutional neural network (CNN), a support vector machine (SVM), and logistic regression were trained and tested using a diverse dataset of 455 images capturing peppers at multiple stages of maturity. The algorithms categorized the crops into binary quality classes (\"good\" vs. \"bad\") by analysing extracted physical attributes, including dimensions, geometry, colouration, and surface uniformity. Model performance was quantified using standard metrics, namely precision, recall, F1 score, and overall accuracy. The findings reveal that while all classifiers achieved acceptable benchmarks, the deep learning CNN model demonstrated superior capabilities, reaching an accuracy of 97%—outperforming logistic regression at 80% and the SVM at 78%. Ultimately, these outcomes validate the integration of automated image analysis systems into modern agricultural pipelines, offering a robust and scalable substitute for manual quality control.","苏格兰帽椒（Capsicum chinense）是全球广泛种植的经济作物，因其独特的烹饪用途和高营养价值而备受青睐。然而，传统的人工检测方式主观性强、速度慢且劳动强度大，难以满足消费者对优质产品日益增长的需求。为解决这一问题，本研究提出了一种基于计算机视觉的框架，旨在实现辣椒分级过程的自动化。研究使用包含455张图像的数据集对卷积神经网络（CNN）、支持向量机（SVM）和逻辑回归进行了训练和测试，该数据集涵盖了辣椒多个成熟阶段的图像。这些算法通过分析提取的物理属性（包括尺寸、几何形状、色泽和表面均匀度），将作物分为二元质量类别（“优质”与“劣质”）。模型性能采用标准指标进行量化评估，即精确率、召回率、F1分数和总体准确率。研究结果表明，尽管所有分类器均达到了可接受的基准水平，但深度学习CNN模型表现更为优异，准确率达到97%，优于逻辑回归的80%和支持向量机的78%。最终，这些结果验证了自动化图像分析系统融入现代农业流程的可行性，为人工质量控制提供了一种稳健且可扩展的替代方案。",null,"ABUAD Journal of Engineering Research and Development (AJERD)","2026-09-12T00:00:00Z","论文",10,false,71,{"impact":17,"substance":18,"depth":19,"authority":20,"freshness":21,"relevant":22,"comment":23},14,20,16,12,9,1,"基于CNN的辣椒品质自动分级研究，97%准确率且对比SVM与逻辑回归，方法清晰、结论可靠，对农产品智能分选有参考价值，但属细分作物应用，影响范围有限。",[25],{"name":10,"url":6},[27,28,29,30,31],"智慧农业","农业人工智能","计算机视觉","辣椒","农产品分级",0,"10.53982\u002Fajerd.2026.0903.06-j",{"doi":33,"openalex_id":35,"authors":36,"venue":10,"cited_by_count":32,"oa_url":6,"card":43,"direction":47,"ingested_from":49},"W7212404580",[37,39,41],{"name":38,"orcid":9},"Babatunde Olayinka Oyefeso",{"name":40,"orcid":9},"Rotimi Rufus Dinrifo",{"name":42,"orcid":9},"Oladapo Oyedeji",{"tldr":44,"method":45,"finding":46,"direction":47,"opportunity":48},"用图像处理和CNN\u002FSVM\u002F逻辑回归对苏格兰帽椒进行好坏二分类质量分级。","455张多成熟度辣椒图像，提取尺寸、几何、颜色、表面均匀性特征，训练CNN、SV","CNN准确率达97%，显著优于逻辑回归80%和SVM78%，验证自动图像分析可替代人工质检。","农业人工智能与决策模型","可扩展到多等级细粒度分级、多品种泛化及田间实时轻量化部署研究。","openalex","2026-09-13T23:30:43.649124Z"]