Computational protocol: Rapid, automated detection of stem canker symptoms in woody perennials using artificial neural network analysis

Similar protocols

Protocol publication

[…] All the images were captured using a SLR camera (Canon EOS 1000D) with 53 mm focal length and 1/15 s exposure time. Two 60 W incandescent light bulbs were used to illuminate the samples from each side. The distance between the lens and the samples was 35 cm. Due to the high resolution of the imagery device (3888 × 2592 pixels), three shoots were placed on a spectralon white platform (SphereOptics) and imaged together in order to enhance the contrast between the foreground and background. The images were captured using EOS utility software (Canon) and saved as JPG files. Individual shoots were cropped from each image and due to small variations in the size of shoots, the resolution of the images varied from 43 × 754 to 282 × 839 pixels. All the images were saved and processed on a Dell desktop computer (Intel® Xeon(R) CPU X5560 @ 2.80 GHz × 16). The automated image analysis software was written in C++ [] utilising the OpenCV Library [] on an Ubuntu 14.04 operating system. [...] ImageJ [] was used to manually measure the disease severity on an image-by-image basis. Firstly, the three cherry shoots were cropped from the original image and converted from RGB to HSI colour space. A threshold was manually chosen to determine the total number of pixels in the shoot (compared to the total in the whole image containing the background). The total number of pixels in the shoot was named R1. The second threshold on the hue channel was used to segment the diseased and healthy areas. As the diseased area always showed darker intensity than the healthy area, the background could be easily separated. The total number of pixels in the diseased area was called R2. The proportion of the diseased area was calculated using the ratio of the diseased area (R2) to the total shoot area (R1). […]

Pipeline specifications

Software tools OpenCV, ImageJ
Application Microscopic phenotype analysis
Organisms Pseudomonas syringae, Prunus persica, Equus caballus
Diseases Bacterial Infections