India's primary industry, agriculture, accounts for the majority of the country's GDP. Weed growth is one of the primary factors that lowers agricultural productivity. In the field, the combination of machine vision and the image processing is showed to be a reliable way for doing point to point real-time detection of weed and for other crop detection, providing with the accurate data to manage the weed tailored to individual sites. The developments in information and electronic technologies are to blame for this. Manual herbicide spraying is one of the most common ways to keep weeds under control. However, this strategy has a lot of negative effects on both agricultural and consumer health. This report summarized the advances in weed detection using open-CV and image detection methods. A detailed explanation was given of the five primary weed identification techniques: firstly preprocessing, secondly feature extraction, third would be segmentation, fourth is extraction, and then lastly classification. The primary difficulty with this model is distinguishing crops from weeds, which can occasionally have identical features and appearances. The model was trained with a sizable dataset in order to address this problem. As a result, this technology will now detect the weed and spray the herbicide in the adequate amount in a targeted manner. This research paper will provide a comprehensive study of various deep learning methods that can be useful for automatic detection and various image processing methodologies. It also provides performance metrics, such as accuracy, and precision for various techniques employed.