提出了一种改进的 DNA遗传算法, 以解决遗传算法用于图像分割时收敛速度慢、易早熟的缺点。利用碱基互补的 DNA编码方式增加种群多样性, 防止陷入局部极值; 设计了基于 DNA分子操作的置换自适应交叉算子和密码子变异算子, 从而提高遗传算法的搜索能力, 有效加快了算法的收敛速度和效率, 并将此算法用于寻找二维 Arimoto熵的最佳阈值, 最后对图像进行分割。普通图像和医学图像的分割结果验证了改进 DNA遗传算法用于图像分割的有效性。

A type of improved DNA genetic algorithm is proposed, which is used to solve the shortcomings like slow rate of convergence and easy precocity. The algorithm has increased the population diversity by using DNA encoding mode of the base complementarities, avoided falling into the local extremum. The replacement adaptive crossover operator and codon mutation operator are designed based on the DNA molecule operation, so it improves the genetic algorithm's search capability, and increases the convergence speed and efficiency. The algorithm is adopted to seek the best threshold of two-dimensional Arimoto. At last, the image is segmented. It is proved that this improved DNA genetic algorithm has improved segmentation validity by both the ordinary image and medical image.