Pests and disease can cause damage to greenhouse crops, significantly reducing crop yield. Diagnosing early symptoms is key to preventing severe harm and the spread to neighboring plants. Powdery mildew is a fungus that can completely cover leaf surfaces, hindering photosynthesis, and proliferate quickly in greenhouse environments. This thesis develops a system of detecting powdery mildew in a greenhouse environment using machine vision. The system includes a specially de signed data acquisition platform. Testing on field images from several greenhouses show excellent detection rates of 85%.