基于小波变换的工具痕迹图像识别研究

Tool Mark Image Recognition Research based on Wavelet Transformation

工具痕迹是一种出现在刑事案件中的重要物证,同时也是串并同类案件的重要依据。充分利用案件现场的工具痕迹有着十分重要的意义。根据小波变换原理,从纹理分析的角度出发,对工具痕迹图像进行二维小波变换,提取灰度共生矩阵的相关特征参数构造BP神经网络对痕迹图像进行分类识别,实验表明,该方法能细致的描述工具痕迹图像的纹理特征,可以辅助刑事技术人员进行工具痕迹图像特征的比对分析。

Tool mark is a kind of material evidences appearing in the criminal cases,at the same time it is an important basis for combining cases.Full use of tool mark is of great significance.According to wavelet transform theory,this paper uses two-dimensional wavelet transform on tool mark images from the perspective of texture analysis,obtains the characteristic features of gray level co-occurrence matrix,designs BP neural network for image classification and recognition.Experimental results show that the method can describe texture features of tool mark images effectively,and also it can support criminal technicians to carry out comparative analysis on tool mark image features.