| Work Detail |
Scientists have developed a novel method that uses live video images to detect shadows on solar panels. It uses computer vision techniques such as gamma transform and histogram matching, resulting in performance that is said to be better than conventional techniques, especially in large installations.
A Chinese research team has developed a novel real-time shadow detection method for photovoltaic modules. It uses computer vision for the task, with techniques that enhance live feed information to identify shadow under changing lighting conditions.
“This method has important reference value for shadow monitoring of large PV arrays. It provides new technical means for intelligent and accurate operation and maintenance of PV systems,” the researchers say. “We believe that this work can potentially monitor the shadow of PV arrays in real time, and we look forward to further research to extend it to more application scenarios.”
Starting from a live video, the novel method analyses each frame, first extracting the surface of the photovoltaic module. Then, to address the problem of illumination variation, the technique uses gamma transformation, which is a computer vision method based on an algorithm that corrects the brightness of the image without references.
In the next step, the image contrast is enhanced using histogram matching, another computer vision technique that brightens an image by looking at the histogram of a reference image. It then uses gray level cutting to segment the shaded portion of the PV modules, from which the actual shaded output is obtained.
“To better simulate the actual working environment, the validation experiment of this paper was designed as a fixed-position chamber, with the environment around the PV module,” the group explains. “The image capture device consists of a mechanical stand, a camera with a frame rate of 30 FPS, and an adjustment device. The PV module consists of 36 solar cells of 220 mm x 770 mm, arranged in 4 columns and 9 rows, glued on the white backsheet and covered with tempered glass.”
The experimental setup was located in Harbin, China, throughout August 2023 and captured 90 videos of one hour each. In total, 4,815 minutes of video were captured, representing different lighting conditions and occlusion levels. About half of the recorded time was used to analyze and train the model, while the rest was used for testing. It was also compared with four other shadow detection models, namely Canny edge detection, multi-level thresholding, random forest, and convolutional neural network (CNN).
“The average recognition accuracy (ACC) of this method is verified to be 0.98 by the test set, which is superior to the existing Canny edge detection recognition method,” the results show. “The F0.5 and F2 values ??of the method are 0.87 and 0.85, respectively, which is good in terms of precision and recall. Moreover, the average time needed by the method to process one image frame is 0.721 s, which is a good real-time performance.”
For compression, the Canny method had an ACC of 0.95, F0.5 of 0.76, F2 of 0.79, and a frame processing time of 0.684 s. The multi-level threshold obtained an ACC of 0.94, F0.5 of 0.75, F2 of 0.77, and a frame processing time of 0.934 s, while the random forest had an ACC of 0.89, F0.5 of 0.31, F2 of 0.35, and a frame processing time of 1.067 s. Finally, the CNN had an ACC of 0.82, F0.5 of 0.38, F2 of 0.62, and a frame processing time of 0.603 s.
Their findings were presented in “ The real-time shadow detection of the PV module by computer vision based on histogram matching and gamma transformation method,” published in Scientific Reports . The research team consisted of scholars from Northeast China Agricultural University and Shenzhen University. |
