It has great significance to use visual inspection methods to do research on energy-saving system. This paper use visual inspection method to detect classroom seating area, illumination, object, and staff numbers to realize classroom-lighting energy-saving system. First, according to the appearing pixel in the photo of classroom background, this paper makes rough marks on seating area, then using horizontal and vertical projection to make accurate marks in order to divide seating area properly. Second, uses the principle of Gray scale images contain illumination information to detect illumination. Third, uses background subtraction method to achieve the segmentation of foreground from background, and different background model is used in different periods. Then improved erosion and expansion techniques are used for denoising color image. Last, on the basis of classic haar classifier algorithm to mark skin and hair in different colors based on skin model and hair color information. An improved connected labeling algorithm is presented to realize statistics on the number of people. Experiment results demonstrate that the system has greater accuracy, higher practical value.