Image-based flame detection (ibfd) algorithms are effective in flame detection, and their advantages and disadvantages directly affect the accuracy and timeliness of flame detection. However, based on the existing res...
详细信息
Image-based flame detection (ibfd) algorithms are effective in flame detection, and their advantages and disadvantages directly affect the accuracy and timeliness of flame detection. However, based on the existing research in China, the ibfd algorithm still lacks a clear standard for distinguishing advantages and disadvantages. This study adopted a fuzzy comprehensive hierarchical analysis method to establish an evaluation method for ibfd and to define the rating standards for 29 third-level evaluation indexes. A three-level index system was installed for algorithm evaluation, and the masses of needles at all levels were determined. An index system was used to assign indexes to the two flame identification algorithms in the YOLOv3 series. The algorithm evaluation was completed according to the mass of the algorithm index, which indicated the reliability and validity of the algorithm evaluation method. Therefore, this study established a set of methods for the performance evaluation of ibfd and thus laid a theoretical foundation for improving relevant standards for image-based flame detectors.
暂无评论