A Win-Win in Precision Ag
Deep learning techniques for precision agriculture enable the optimization of management practices, including water and chemical applications, benefiting both the farmer and the environment. We’ll explore two use cases. First, we collect high-resolution aerial imagery and use a deep learning based density-estimation approach to count and localize flowering pineapple plants across a field, enabling precision application of chemicals and reducing waste. Second, we use longitudinal aerial imagery of corn and soy fields to detect and predict nutrient deficiency stress. By leveraging a U-Net and Convolutional LSTM, we are able to detect and predict stress up to three weeks earlier.
Key Takeaways: Advances in deep learning and high-resolution image acquisition are revolutionizing precision agriculture Deep density-estimation techniques enable us to count millions of plants from aerial images in just a few seconds Incorporating the temporal element of our data enables us to do better detection and prediction of key issues in the field like identifying plants under stress
Jennifer Hobbs is the Director of Machine Learning at IntelinAir, an ag-tech startup using computer vision and machine learning to deliver intelligence and insights to the agriculture industry. Her team is responsible for the development and delivery of models which identify relevant patterns in the field and alerts users to them. She completed her PhD in Physics and Astronomy at Northwestern University. Throughout her career she has been involved in all phases of the machine learning lifecycle, transforming raw data into compelling technology products through data modeling and architecture, pipeline design and management, machine learning, and visualization.