Role of IoT, image processing and machine learning techniques in weed detection: a review

Weed is one of the main reasons that cause damage to the crop. While cultivating the crops, many of the farmers are facing problems due to weed growth. The problems are the damage of the crop, the pest and diseases, cultivation cost, environmental pollution due to herbicides and the time-consumed to remove the weed. Machine learning, IoT and image processing are promising domains, which can be used to detect and control the weed from the agricultural land. In this article, different types of cameras were used to collect and compare the images; pre-processing techniques were used to remove the noise; all the segmentation methods, feature extraction, feature selection, training and testing methods were used to identify the weed from the crop. Aside from weed identification, the different controlling methods used to control weed from the agriculture field were also discussed. This review article presents all the information related to IoT, machine learning and image processing to detect weeds. Finally, this article describes the different challenges of weed detection and possible solutions to each challenge.