Detection of Anomalies in Citrus Leaves Using Digital Image Processing and T2 Hotelling Multivariate Control Chart

Citrus is one of the important horticulture products in Indonesia. Production fluctuations that occurred in several production centers were caused by the attack of the disease. Usually, identification and detection of citrus plant anomalies were carried out by observing symptoms of disease in leaves directly on the field. To confirm anomalies of citrus plants caused by disease, several tests could be carried out, including physiological assessment and testing their genetic properties. To use those techniques as early warning technique to detect citrus plant anomalies is not efficient, because the test takes time and costs. The technology that is currently developing as an alternative to a conventional method for early warning observation is image processing. RGB information extracted from image processing was used by T
²
Hotelling multivariate control chart to detect changes in citrus leaf color. T
²
Hotelling multivariate control chart uses RGB information as multivariate input to determine occurred anomalies. The application of image processing technique and T
²
Hotelling multivariate control chart could help early examination of the signs of citrus plants anomalies that are likely caused by a disease. T
²
Hotelling multivariate control chart method was able to determine between healthy and anomalies occured citrus leaves with accuracy 83%.