I think a machine learning algorithm wouldn't care about that, because with a large enough training data set it would start to account for that and be able to accurately predict energy output based on the image alone.
Another issue with this dataset is the overlay changing over time in text content, font, and colour. The algorithm might overfit and think e.g. yellow font presence means higher output simply because the output was higher during that period. You could strip away the text, but then you're introducing potential errors into the dataset yourself.