Machine Learning Methods for Weed Recognition in Corn Fields: A Review

Herbicide use is a common practice for managing weeds in wheatgrass fields, but it can be costly, raise ecological concerns, and lead to herbicide resistance. A potential solution to this problem is using machine learning models for precise weed identification. This study provides an overview of the key ML techniques utilized in wheatgrass weed identification, including classifying and detecting objects. A performance evaluation measure, such as accuracy in classification and score F1, were also discussed. Furthermore, potential areas for future research are highlighted, such as increasing a data capacity through enhancing data, leveraging passing learning, and enhancing comprehension of artificial neural networks to avoid excessive fitting and boost transparency. Typically, digital images are utilized as input data in ML weed identification, although hyperspectral data is sometimes employed. The majority of current studies utilize support vector machines and neural networks for this purpose.