Medicinal plants are essential to healthcare since ancient times and are integral to developing drugs and other medical treatments. More than 25% of medicines in developed countries are produced from medicinal plants, while in developing countries, approximately 80% of individuals receive primary healthcare from these plants. Traditionally, these plants are identified manually by experts, which is tedious, time-consuming, subjective and dependent on the availability of experts. Furthermore, a wrong detection can result in serious health issues or death. This signifies the need for a more reliable approach to identifying medicinal plants, which is accurate and practical. Several automated methods were proposed previously, utilizing deep learning and traditional machine learning (TML) techniques, but they require singular leaf images and failed to achieve sufficient accuracy when demonstrated in a different setting. Capturing singular leaf images for each plant is also time-consuming and laborious. This paper presents a robust, accurate and practical system to identify medicinal plants from smartphone-captured plant images in the site of plants. The proposed system utilized a cascaded architecture to extract features using a pre-trained ResNet50 model, which were optimized using Particle Swarm Optimization (PSO) to classify the plants using a Support Vector Machine (SVM). The proposed ResNet50-PSO-SVM network classified seven medicinal plants with 99.60% accuracy, outperforming the state-of-the-art (99%). The system was demonstrated for three different smartphones, classifying an image in 0.15 seconds with 97.79% accuracy on average. The system’s high accuracy, rapid identification time and robustness ensured its practical use.

The “WaterHyacinth” dataset, a recently gathered collection of images featuring four distinct species of Water hyacinth from different regions of Bangladesh, is presented in this article. There are four different classifications: Lemna minor, Eichhornia crassipes, Monochoria korsakowii, and Pistia stratiotes. The collection consists of 1790 original images and in addition 4050 augmented photos of Water hyacinth species. Every original picture was captured with the appropriate background and in sufficient natural light. Every image was correctly placed in its corresponding subfolder, providing optimal use of the pictures by various machine learning and deep learning models. Researchers could make major progress in agriculture, environmental monitoring, aquatic science, and remote sensing domains by utilizing this enormous dataset and various machine learning and deep learning approaches. In addition to opening opportunities for significant developments in these domains, it offers an essential asset for further study.

This study presents a recently compiled dataset called “BDHusk,” which encompasses a wide range of husk images representing eight different husk species as a component of cattle feed sourced from different locales in Sirajganj, Bangladesh. The following are eight husk species: Oryza sativa, Zea mays, Triticum aestivum, Cicer arietinum, Lens culinaris, Glycine max, Lathyrus sativus, and Pisum sativum var. arvense L. Poiret. This dataset consists of a total of 2,400 original images and an additional 9,280 augmented images, all showcasing various husk species. Every single one of the original images was taken with the right backdrop and in enough amount of natural light. Every image was appropriately positioned into its respective subfolder, enabling a wide variety of machine learning and deep learning models to make the most effective use of the images. By utilizing this extensive dataset and employing various machine learning and deep learning techniques, researchers have the potential to achieve significant advancements in the fields of agriculture, food and nutrition science, environmental monitoring, and computer sciences. This dataset allows researchers to improve cattle feeding using data-driven methods. Researchers can improve cattle health and production by improving feed compositions. Furthermore, it not only presents potential for substantial advancements in these fields but also serves as a crucial resource for future research endeavors.

We introduce the “Ripen Banana” dataset, a newly developed collection featuring two distinct classes of ripen banana images: carbide and non-carbide. The dataset contains images from raw to ripe bananas that have been ripened with carbide and without carbide. The images collected from various locations in Sirajganj, Bangladesh. It contains 1404 original images and 6410 augmented images, resulting in 7814 images overall. The original images were taken against carefully selected white backgrounds. This was made it by evaluating these images for their sharpness and brightness levels to ensure they meet the quality standards that are required. After that, we organize each image into specified subfolders to ensure simple and fast access for large range of machine learning and deep learning models. This dataset offers significant opportunities for advancements in agricultural practices, nutrition and the food chain, food environment monitor, and computer vision technologies. This “Ripen Banana”