This study uses an image-processing technique to determine the damage level caused by two-spotted spider mite (Tetranychus urticae Koch) to cucumber plants and changes in the number of mites in a greenhouse. Firstly, a new agricultural platform was developed to ensure camera stability for capturing quality images. The images of 50 leaves infested with T. urticae were captured weekly for 5 weeks with the platform, which resulted in 250 images. Fifty of these captured images were randomly selected and processed with an image-processing algorithm developed using an image processing toolbox module of MATLAB. The results obtained from the image processing algorithm were compared with expert observations. The image-processing method predicted the damage with 3.91 root mean squared error (RMSE). A highly significant positive relationship was found between image processing and expert observations. The results indicate that this new image-processing method may be successfully used in place of expert observation to determine T. urticae damage in greenhouses.
Keywords: Agricultural platforms; Agricultural robots; Cucumber; Image processing; Two-spotted spider mite.