The majority of crop vision datasets include domain-specific annotations for use in advanced deep learning applications in both offline and online agricultural settings. In this study, we present BellCrop, a novel dataset of 4,860 high-resolution images of bell pepper leaves divided into three categories: Healthy, Powdery mildew-infected, and Magnesium-deficient leaves. Unlike previous datasets, which were primarily collected in controlled environments, these images were captured in greenhouses under natural conditions. Images were taken with a camera with a resolution of 3468 × 4624 pixels. Deep learning-based classifiers such as VGG19, Xception, and ResNet50 were used to detect these diseases during the early stages of plant growth. Despite the challenging nature of the dataset, the results show detection accuracies of 87%, 81%, and 80%. BellCrop study demonstrates the potential of deep learning models in automating the early detection of leaf diseases in bell peppers. It has significant implications for precision agriculture, allowing for more timely interventions and improved crop management efficiency. Additionally, the study looks into the use of advanced models known as Transformers. EfficientNet-B7 and Vision Transformers (DeiT) achieved more than 90% accuracy rates, improving the reliability of disease classification in agricultural settings.

This review paper discusses the usage of Computer Vision (CV) and Machine Learning (ML) in greenhouse environments for crop grading. It focuses on the progress of autonomous quality gardening using crop image acquisition, preprocessing, and grading over the past five years. The review focused on greenhouse crops: apple, mango, bell pepper, papaya, tomatoes, strawberry, carrot, and chilli. The study focuses on noise reduction, image segmentation, and feature extraction techniques. To improve the assessment and classification of crop quality. Classification using ML algorithms, such as Random Forest, Support Vector Machines (SVM), and K-nearest Neighbours (KNN), are discussed. The addition of AI-driven approaches in agricultural practices is prioritized to enhance crop quality, minimize environmental impact, and satisfy market standards. This review is a comprehensive resource for researchers and practitioners interested in utilizing AI technologies to achieve sustainable and efficient crop production.

This paper presents a machine learning based, cost effective approach to accurately identify the Nitrogen fertilizer demand in paddy fields based on the leaf color. A smart mobile phone mounted on a UAV (Unmanned Aerial Vehicle) was used to capture aerial images of different areas in paddy fields. The proposed approach was used to decide the Nitrogen fertilizer demand for each captured region based on the overall color level identified for the corresponding image, with the help of the color levels and the fertilizer recommendations specified in LCC (Leaf Color Chart). Evaluation results reveal an overall accuracy of 86.5% for the predictions of LCC color levels. Average time taken to predict a single instance was 6.64 seconds.

In some developing countries, widows are looked down upon and are often considered inauspicious especially in rural regions. Some societies even consider them and their issues invisible. This paper presents findings from a qualitative study focused on understanding how technology could facilitate entrepreneurial and DIY activities of widows from rural Sri Lanka. We conducted semi-structured interviews and field observations with thirteen widows from low socio-economic backgrounds, who were involved in various small-scale entrepreneurial activities. Our findings showed three central aspects associated with their entrepreneurial activities which can be supported through technology: initial stages of entrepreneurship, balancing work with life, and dealing with exploitations. This paper explores how gender inequality in a social context affects marginalized women in rural Sri Lanka in conducting their entrepreneurial efforts. In particular, we highlight resilient practices that the participants apply to support their entrepreneurial activities. With an "assets-based approach" we conclude by providing implications for policymakers, media, and HCI practitioners to support this inbuilt resilience by leveraging their current assets.

Microfinance is an inherently social process. Over the years, it has emerged as an essential means for providing financial services to the "underbanked" population in developing countries. This paper presents a qualitative study focused on understanding existing cooperative practices associated with microfinance in rural Sri Lanka. Through semi-structured interviews, group discussions and visits to microfinance centres, we found that microfinancing involves much more than financial transactions, rather it is strongly habituated in the sociocultural fabric of communities. Our findings show that three factors affect the cooperative process of microfinancing: trust and credibility, community support, and familial assistance. Using examples from the field, this paper discusses the importance of these factors in microfinance activities and contributes an elaborated account of cooperative practices that support microfinance processes. We also highlight how the local sociocultural structures and practices shape the way microfinance processes are being handled. We contribute towards an in-depth understanding of such practices that can be useful for microfinance institutions (MFIs) and technology designers. Consequently, we advocate for an intermediary path for technology interventions where technology and people work together rather than technology replacing people.