Genetic Algorithm is a macro-heuristic algorithm based on the theory of natural selection. It uses genetic operators such as selection, crossover, and mutation to generate or search for higher-quality solutions in the population, so as to achieve the optimal solution to the problem.

 

Genetic algorithm can realize thermal infrared image enhancement well, and its essence is to find a gray scale transformation relationship. The genetic algorithm first encodes the gray-scale correspondence between the output image and the input image, and initializes the gray-scale correspondence (that is, the first generation of individuals) randomly, then constructs a fitness function according to a certain image quality evaluation standard, and then repeatedly uses the genetic operation (Including mutation, crossover, selection) Evolution generates new offspring individuals until the optimization criterion is met, thus finding the optimal or nearly optimal gray-scale transformation relationship, and finally applying this relationship to achieve image enhancement.

 

Constructing a reasonable fitness function is the key to obtain high-quality solutions by genetic algorithm. In addition, the genetic algorithm should pay attention to solve the problem of excessive calculation, because it is possible to set a larger population size for the problem and go through a longer evolutionary process to search for the solution.