Computer Vision Based Detection of Weed in Ginger and Sugarcane Crop for Automated Farming System

In an agricultural field the machine learning is proven to be work effectively in an efficient crop management the weed detection in agricultural fields is a critical duty. Manually weed detection with the help of labor is unsuitable and requires more time in large-scale agriculture. As a result, the demand for automated machine learning -based weed detection systems is rising. The presented in the research Sugarcane and Ginger Crops and their repetitive weed is considered for the classification. In recent years, weed detection has been automated using computer vision techniques including image processing using of feature descriptors and machine learning algorithms. This research papers gives a focus on weed detection in agricultural settings. The paper discusses numerous weed identification techniques, such as Image pre-processing, feature description using BRISK, SIFT. The Novel method of feature descriptor is also discussed which is Fusion of BRISK and SIFT. The PCA and machine learning classifiers such as Random Forest, Decision tree, Gaussian Naive bayes as well as KNN are applied. Each method is discussed in the paper, as well as how well it performs in certain situations. Approaches for weed detection are a promising field of study that can assist farmers in lowering labor costs, raising crop yields, and reducing the usage of herbicides.