<p>In computer vision, Object detection is a hot research topic, which detects objects in images. For effective object detection, more texture details are provided by the visual and infrared image fusion methods. In recent years, several methods have been developed for image fusion and object detection. However, the existing methods fail to preserve the significant structural and textural information, resulting in subpar performance, and also face computational complexity problems. To overcome these drawbacks in existing methods, this research proposes the Prairie Dog Search Optimization enabled Enhanced Triplet Spatial Channel attention-based Generative Adversarial Network (Pr-ETSC-GAN) to effectively detect the objects and enhance the image fusion. The Pr-ETSC-GAN model integrates the Prairie Dog Search Optimization (PrSO), which enhances the hyperparameter tuning, and the ETSC attention helps the proposed model to significantly focus on the salient information. The proposed model effectively fuses the image and detects the object with low computational complexity, and increases the performance. Extensive experiments show that the Pr-ETSC-GAN model achieves excellent results, reporting a high accuracy of 97.92%, precision of 98%, Intersection over Union (IoU) of 0.87, recall of 97.95%, mAP of 0.93, and F1 score of 97.95% on training with the CAMEL dataset. Further, the Pr-ETSC-GAN model achieves an accuracy of 96.93%, precision of 97%, IoU of 0.87, Recall of 96.94%, F1 Score of 96.93%, and mAP of 0.92 for K-Fold analysis.</p>