These techniques enable cameras and closed-circuit television (CCTV) systems to be used for smoke and fire detection, thus providing remote coverage for wider areas. To overcome these limitations, video-based fire detection systems are currently commonly used mainly supported by new emerging image processing and computer vision techniques. Moreover, such devices can be easily damaged by the smoke and high temperatures developed during a fire. Traditional smoke detector devices are able to identify the presence of smoke only in the close proximity of the source of emission, but they lack the ability to signal the presence non-local hazards. Fire protection and prevention systems are available in the majority of public buildings and they are ubiquitous in public and private transportation. Prevention of fire accidents is an important safety, economic and environmental issue that is constantly addressed in various research fields. Finally, we developed a portable solution for embedded systems and evaluated its performance for the Raspberry Pi 3 and the Nvidia Jetson Nano. The proposed technique has been compared to the state of the art methods available in literature by using several videos of public and non-public dataset showing an improvement in the metrics. The algorithm provides also an alarm generator that can trigger an alarm signal if the smoke is persistent in a time window of 3 s. In the final step the convolutional neural network verifies the actual presence of smoke in the proposed regions of interest. This technique allows the automatic selection of specific regions of interest within the image by the generation of bounding boxes for gray colored moving objects. The algorithm combines a traditional feature detector based on Kalman filtering and motion detection, and a lightweight shallow convolutional neural network. In this work presents a hybrid approach to assess the rapid and precise identification of smoke in a video sequence. Including intelligent video-based techniques in existing camera infrastructure enables faster response time if compared to traditional analog smoke detectors. Smoke detection represents a critical task for avoiding large scale fire disaster in industrial environment and cities.

Realflow smoke