摘要
来自多传感器的目标特征往往是高维数的,并且包含了更多的冗余信息和噪声。为了减小数据获取的代价,提高目标识别器的性能和效率,提出了基于遗传算法(GA)的多传感器目标识别系统特征优化方法。将遗传算法与神经网络目标分类器结合,通过识别结果的反馈信息,控制GA的遗传进化方向,从而实现特征优化。为了克服遗传算法的未成熟收敛问题,提出了相关选择与自适应遗传算子相结合的改进遗传算法。仿真实验结果验证了方法的有效性。
The features of target from multi-sensor system are generally high dimensional, redundant and noisy. A genetic algorithm (GA) based feature optimization approach was proposed for multi-sensor target recognition system to reduce the cost of acquiring data and improve the performances and efficiency of recognizer. Incorporated a neural network classifier, the evolution of GA was directed to optimization with the information feedback. Since a standard GA has the shortage of premature convergence, an improved genetic algorithm was designed to prevent it. The simulated experimental results for the feature optimization show that the proposed method is effective.
出处
《光学技术》
EI
CAS
CSCD
北大核心
2005年第3期420-423,426,共5页
Optical Technique
