Dust, fallen leaves, or obstructions on solar panels are automatically identified as hazards by OpenMV.

A photovoltaic panel that “appears to be fine” does not necessarily mean it is operating at normal power generation efficiency. Localised shading and dust accumulation often persist for extended periods yet remain difficult to detect. OpenMV enables photovoltaic operations and maintenance to move beyond full reliance on manual inspections.

During the operation of photovoltaic power plants, foreign matter such as dust accumulation, fallen leaves, and bird droppings on the surface of photovoltaic panels directly reduces light absorption efficiency, leading to decreased power generation; over the long term, this may even accelerate component ageing and shorten equipment service life.

For large-scale photovoltaic arrays, manual inspection is not only inefficient but also unable to promptly detect local anomalies. Leveraging SingTown’s OpenMV intelligent cameras, long-term and stable visual monitoring of photovoltaic panel operations can be achieved, enabling automatic identification and timely resolution of foreign object issues.

The camera can be securely mounted on a support structure to cover multiple photovoltaic (PV) module areas and continuously capture images of the PV panel surfaces. OpenMV performs region segmentation at the edge to delineate the effective operating areas of the PV panels, and automatically detects anomalies such as abnormal dust accumulation, leaf obstruction, or local contamination by combining foreign-object feature recognition with status comparison analysis.

By comparing the current image with a normal reference, the system identifies abnormal areas affecting power generation efficiency and outputs the detection results. These results can directly trigger cleaning equipment, operations and maintenance systems, or alert modules at the edge, enabling immediate action upon detection without uploading large volumes of image data to the cloud.

Leveraging OpenMV’s edge computing capabilities and low-power design, this solution is especially suitable for long-term deployment in large-scale, distributed photovoltaic scenarios.