基于改进视觉注意机制的跑道异物检测算法

Foreign Object Debris Detection Based on Improved Visual Attention Mechanism

机场跑道异物是影响航空器安全的重要因素。将人类视觉注意机制引入到基于视频跑道异物检测中,在传统Itti模型基础上,进一步提取纹理、边缘特征,依据显著图能量对各个特征进行加权求和,获得最终的视觉显著图。实验结果表明,方法合成的显著图克服了传统Itti模型的不足,对跑道异物的检测效果明显提高。

FOD（ Foreign Object Debris） is an important factor that affecting the safety of aircraft. The human visual attention mechanism is introduced into the video-based foreign object debris detection. Based on the traditional Itti＇s model,the texture and edge features are further extracted to form the final visual saliency map by image energy. The experimental results show that,this method can overcome the shortcomings of the traditional Itti＇s model,and improve the detection of foreign object debris.