Early detection of stress in strawberry plants using novel image analysis techniques

This project aims to improve the detection and prediction of events such as drought or disease stress before associated damage and financial costs are incurred. To achieve this, cutting edge imaging technology and software algorithms will be used. To replicate this process manually would require heavy investment in terms of manual inspection time and training. The additional advantage of developing technology based approaches is the removal of the subjectivity present in observations made by an individual person, and the ability to automate this process in the future, further reducing labour costs and improving efficiency. The imaging equipment used can “see” regions of the visible spectrum humans cannot, increasing further the detection potential of the approach.

To achieve this, the project will investigate the use of hyperspectral* imaging to identify differences in growth in glasshouse-based strawberry crops. Time-series hyperspectral datasets will capture crop growth as a series of images. The hyperspectral image acquisition system will be used at East Malling Research (EMR) in a glasshouse, and will capture time lapse images of multiple strawberry plants. The resulting image data will be analysed using novel image analysis techniques to extract information about the plants form the images. The extracted data will be a valuable resource documenting crop growth, and data mining approaches can then be used to pick out significant parts of the dataset to enable detection of visible effects in the crop, and ideally predict the onset of effects before the crop is damaged.