Big data and machine learning for crop protection

•This paper presents a review of some significant research efforts in crop protection using Big data with the focus on weed control and management .•Some machine learning techniques for Big data analytics are also reviewed.•The potential of using Markov random fields (MRF) for herbicide resistance modeling of ryegrass is explored.•Experiments and data analytics have been performed on data collected from farms in Australia.•Results have revealed the good performance of this approach. Crop protection is the science and practice of managing plant diseases, weeds and other pests. Weed management and control are important given that crop yield losses caused by pests and weeds are high. However, farmers face increased complexity of weed control due to evolved resistance to herbicides. This paper first presents a brief review of some significant research efforts in crop protection using Big data with the focus on weed control and management followed by some potential applications. Some machine learning techniques for Big data analytics are also reviewed. The outlook for Big data and machine learning in crop protection is very promising. The potential of using Markov random fields (MRF) which takes into account the spatial component among neighboring sites for herbicide resistance modeling of ryegrass is then explored. To the best of our knowledge, no similar work of modeling herbicide resistance using the MRF has been reported. Experiments and data analytics have been performed on data collected from farms in Australia. Results have revealed the good performance of our approach.