The invention discloses a
machine vision based real-time diagnosis method and
system of vegetable leaf diseases. The
system comprises a
disease image acquiring hardware part and a real-time diagnosis
software part, wherein the image acquiring hardware part consists of a lighting box, a
light source bracket, a
light source, a round hole at the top of a box body, a
digital camera, a camera fixing rod, an optical lens, a
polarizer, a box body side door, an objective table and a computer. The method comprises the following steps: acquiring the symptom image of the
leaf disease of vegetable to be detected; carrying out image treatment and scab division by utilizing real-time diagnosis
software; acquiring the characteristic indexes of scab, such as color, texture and shape; and carrying out real-time diagnosis on the vegetable
leaf disease by utilizing a
disease identification model. The real-time diagnosis method and
system can be directly used for detecting diseases in filed, plastic houses or greenhouses so as to achieve the rapid, stable and real-time diagnosis of various vegetable leaf diseases; no chemical
reagent is used, therefore detection cost is reduced and no environment
pollution is caused. The real-time diagnosis method and system can be well applied to
disease monitoring.