The areas of application of mobile robots are becoming increasingly extensive. Most mobile robots use visual images to obtain information about the surrounding environment, including information about the floor area that can be traversed or obstacles to be avoided. In this study, we propose a multiple-classifier method based on an improved fuzzy integral (IFI) for floor area detection using charge-coupled device sensors. The multiple classifiers include deep learning networks (DLNs) and an area contour detection (ACD) algorithm. The IFI in the proposed method is different from the traditional fuzzy integral in that its fuzzy density value is optimized using the artificial bee colony algorithm. The proposed method takes the outputs of several classifiers and the ACD algorithm as the input. The IFI is used to evaluate each classifier to leverage its advantages for the proposed method. Finally, the classification results are calculated to achieve accurate floor area detection. To verify the performance of the proposed method, we used the public MIT scene dataset and images of various indoor environments as the study dataset, with 200, 48, and 48 images used as training data, verification data, and test data, respectively. The experimental results revealed that the DLN–ACD–IFI combination achieved an average floor area detection accuracy of 97.8%, thus improving the overall recognition rate.