复杂环境下基于视觉显著性特征的铁轨识别方法

Track detection approach in complex environment based on saliency features

为解决复杂环境下铁轨边缘识别问题,提出一种基于显著性特征的复杂环境铁轨识别方法:通过引入多尺度Gabor能量算子和环境抑制算子,建立基于视觉机制的边缘检测模型,实现铁轨边缘特征的检测,并对环境干扰信息进行抑制;建立铁轨显著性评价模型,对检测结果中对非铁轨边缘进行滤除;提出基于统计学的铁轨特征增强方法,对铁轨边缘片段进行连接增强。试验结果表明,该方法对于光照变化和噪声干扰的鲁棒性较强,相比其他的边缘检测方法,更适用于复杂环境下的铁轨识别,同时可以应用于相似检测环境下的其他场景。

A new track edge detection method based on the saliency features was proposed to overcome the detection problem in complex environment.In order to extract the edge of track features,the muti-scale Gabor filter was introduced and the suppressing operator was used to realize inhibition of environmental interference information,which together established the fusion detection model like the Human Visual System.The evaluation model for track salient features was built to further filter out false edges,while enhancing the right edges by angle statistics method.Experimental examples were made including under the condition of varying illumination or noise.The results show that the proposed method is more suitable for track edge detection in complex environment and similar detection scenes,compared with other detection methods.