Improved-Fuzzy-Integrals-based Deep Learning Networks for Floor Area Detection

• Main Authors: Yun-Zhen Xie, 謝昀蓁
• Other Authors: Hsueh-Yi Lin
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/zfn2w9

碩士 === 國立勤益科技大學 === 資訊工程系 === 106 === With the advancement of technology, the technology of mobile robots is becoming more and more widely used. Nowadays, most of visual images are used to obtain information about the surrounding environment. This thesis proposes an Improved Fuzzy Integrals (IFI) method, which combines several deep learning networks, and introduces the Area Contour Detection Algorithm (ACD) to enhance the details to form the new image segmentation architecture. Detect the location of the floor area. The improved fuzzy integral is based on the output of several classifiers. It intelligently evaluates the advantages and disadvantages of each classifier and extracts its advantages. Finally, the classification result of the current input data is calculated to achieve a more accurate floor area location. The experimental results show that the improved fuzzy integral combines different detection methods, which improves the overall recognition rate.