A novel custom optimized convolutional neural network for a satellite image by using forest fire detection

One of the main degradation occur in forest region is fire attack. Till now, million hectares of forest areas are burned out due to the environmental and human impact. This will severely cause a great harm to environment, wild animals and ours. To safeguard all those things, prediction is necessary for avoiding the fire allegation in advance. In this research work, Convolutional Neural Network (CNN) with optimized technique is designed within the layer to improve the accuracy level and reduce the error rate. This new approach is named as custom optimized CNN whereas the objective function of Antlion Optimization (ALO) is inserted inside parametric ReLU of CNN layer. The data source used in this research is Landsat satellite image with considering five different forest fire region for testing, training and validation. The experiment is processed to classify the data with custom optimized CNN which in terms it results the detection of different forest fire images. The obtained output accuracy of five different classes is compared with different other existing techniques like CNN, CNN with custom layer without optimization and MSVM to prove its efficiency.