Fires can cause devastating damage to lands, properties, and humans. Many countries suffer from huge financial losses due to these fires. Therefore, there is a need to implement a practical solution to spot fires effectively and accurately. Deep-learning algorithms and artificial intelligence have been deployed recently in various fields, such as monitoring systems, economics, and detection. This paper proposes a New Light Ensemble Deep-Learning Framework (NLEDLF). This framework consists of two deep-learning technologies, which are a New Generative Adversarial Network (NGAN) and a New Convolutional Neural Network (NCNN). These two tools are incorporated into the framework along with some image preprocessing methods to detect fires using pixels. The proposed framework achieves a reasonable.