An optimal segmentation framework for early crack and fire detection using Potoo swarm optimization algorithm

Abstract

The process of dividing an image into multiple segments or regions with similar color, texture, and intensity is known as segmentation. Simplifying or making the image representation more meaningful and easier to analyze is the goal of image segmentation. For computer vision, it is used, medical imaging, and remote sensing applications for object recognition, tracking, and classification. Segmentation is needed for damage and fire images to identify and isolate the regions of interest in the images. In the case of damage detection, the segmentation can help to identify the areas of the structure that are damaged or compromised. This information can be used to assess the severity of the damage and plan for repairs or maintenance. Similarly, in the case of fire detection, segmentation can help to isolate the regions of the image that contain flames or smoke, which can be used to trigger alarms and alerts for rapid response and fire suppression. Overall, segmentation helps to extract the relevant information from the image, which can be used for decision-making and action planning. This research introduces a new framework for detecting damage and fire early on in images by segmenting them optimally using the Potoo swarm optimization (PTSO) algorithm. To enhance the accuracy of crack/fire detection, we first apply a benchmark preprocessing technique that removes unwanted artifacts and enhances image quality. Then, we design a novel PTSO algorithm inspired by the nocturnal vision and hunting skills of Potoo birds, which specifically targets crack/fire segmentation. We evaluate the performance of proposed PTSO segmentation framework on benchmark datasets and compare its results with existing optimal segmentation frameworks.