Date of Publication :17th February 2024

Abstract: By combining cutting-edge Machine Learning (ML) and Deep Learning techniques, this initiative tackles the vital need for medicinal plant identification in conventional medicine, pharmaceuticals, and conservation. In the realm of automated plant recognition, machine learning (ML) classification algorithms effectively classify medicinal plants according to their traits, whereas convolutional neural networks (CNNs) enable image-based identification. This method helps pharmaceutical companies quickly find bioactive substances, correctly identifies medicinal plants, and supports conservation efforts by identifying and safeguarding endangered species. The main goal of the research is to automate the identification of medicinal plants by utilizing image processing techniques. As part of the process, a broad collection comprising high resolution photographs of different plant components is created. Data quality is enhanced by pre-processing methods including noise reduction and picture enhancement. The process of feature extraction incorporates morphological, textural, and color elements to convert visual traits into quantitative data. The system can identify and distinguish between different plant species thanks to these traits, which are used as input by machine learning (ML) techniques including CNNs and SVMs. The model is iteratively refined to improve resilience after being trained and verified on the prepared dataset. Accuracy, precision, recall, and F1 score are examples of performance measurements that guarantee reliability in a variety of environmental settings and image alterations. In summary, this study at the nexus of cutting-edge technology and conventional wisdom provides an advanced, socially significant method for medicinal plant identification. It enhances medical procedures, supports sustainable relationships with natural resources, and aids in the protection of biodiversity.

Reference :