Combining computer vision and deep learning to enable ultra-scale aerial phenotyping and precision agriculture: A case study of lettuce production

• DOI: 10.1038/s41438-019-0151-5
• Publication Date: 2019-06-01T08:03:49Z
• Authors: Alan Bauer; Aaron George Bostrom; Joshua Ball; Christopher Apple...; Tao Cheng; Stephen Laycock; Sergio Moreno Rojas; Jacob Kirwan; Ji Zhou
• Keywords: Precision Agriculture

Aerial imagery is regularly used by crop researchers, growers and farmers to monitor crops during the growing season. To extract meaningful information from large-scale aerial images collected field, high-throughput phenotypic analysis solutions are required, which not only produce high-quality measures of key traits, but also support professionals make prompt reliable management decisions. Here, we report AirSurf, an automated open-source analytic platform that combines modern computer vision, up-to-date machine learning, modular software engineering in order measure yield-related phenotypes ultra-large imagery. quantify millions in-field lettuces acquired fixed-wing light aircrafts equipped with normalised difference vegetation index (NDVI) sensors, customised AirSurf combining vision algorithms a deep-learning classifier trained over 100,000 labelled lettuce signals. The tailored platform, AirSurf-Lettuce, capable scoring categorising iceberg high accuracy (>98%). Furthermore, novel functions have been developed map size distribution across based on associated global positioning system (GPS) tagged harvest regions identified enable conduct precision agricultural practises improve actual yield as well marketability before harvest.