Bar-type steel is commonly used in engineering facilities, which is made from the raw material of steel wire with high-speed rolling. A hot steel-bar stack (HSBS) accident is a serious accident wherein a hot steel bar flies out from a bar stack fixed on a trolley during manufacturing. If not prevented on time, it can damage production equipment and cause fire and personal injury. At present, the monitoring and identification of HSBS accidents during the rolling manufacturing process are still limited to manual observation. We lack advanced monitoring and identification methods. Finding an effective, accurate, and rapid identification methods as well as a treatment method for detecting an HSBS accident in the rolling manufacturing process is an urgent issue. To solve this problem, we propose a novel three-in-one integral recognition (TIOIR) method based on the bagging and boosting ensemble learning schemes. The TIOIR method integrates the maximum distance positioning, corner detection positioning, and ablation methods to better identify different features of HSBS images. Furthermore, we designed and built a fault diagnosis system of HSBS accident detection, which includes temperature and visual sensors, visual detection devices, and a remote control and computing unit embedded with our proposed TIOIR scheme. Through the operation of the fault diagnosis system, we carried out an actual identification experiment of HSBS accident detection in the rolling field, and the obtained real-time recognition rate was as high as 97%.