基于数据挖掘的光照变化复杂图像识别方法  

Recognition method of complex image with illumination change based on data mining

作　　者：张延军[1] 乔德军[1] 张素蕙[1] ZHANG Yanjun;QIAO Dejun;ZHANG Suhui(Handan University,Handan Hebei 056005,China)

机构地区：[1]邯郸学院,河北邯郸056005

出　　处：《激光杂志》2022年第11期129-133,共5页Laser Journal

基　　金：河北省科学技术厅课题(No.112135128)。

摘　　要：光照变化复杂图像识别一直是图像研究领域中的难点,当前光照变化复杂图像识别方法无法消除光照变化对图像识别结果的干扰,导致光照变化复杂图像识别精度低,为了获得更优的光照变化复杂图像识别结果,设计基于数据挖掘的光照变化复杂图像识别方法。首先收集光照变化复杂图像,采用双边滤波算法对光照变化复杂图像进行预处理,增强光照变化复杂图像清晰度,然后采用双树复小波域变换提取光照变化复杂图像识别特征,并采用数据挖掘方法建立光照变化复杂图像识别模型。实验结果表明,数据挖掘可以描述光照变化复杂图像的类型,光照变化复杂图像识别精度超过95%,有效降低了光照变化复杂图像识别错误率,具有比较明显的优势。The recognition of complex image with illumination change has always been a difficulty in the field of image research. The current recognition methods of complex image with illumination change can not eliminate the interference of illumination change on the image recognition results, which leads to low recognition accuracy of complex image with illumination change, In this paper, we design a data mining based recognition method for complex images with illumination changes. Firstly, the complex image of illumination change is collected, and the bilateral filtering algorithm is used to preprocess the complex image of illumination change, so as to enhance the clarity of the complex image of illumination change. Then, the dual tree complex wavelet domain transform is used to extract the recognition features of the complex image of illumination change, and the data mining method is used to establish the recognition model of the complex image of illumination change. The experimental results show that data mining can describe the types of complex images with light changes, and the recognition accuracy of complex images with light changes is more than 95%, which effectively reduces the recognition error rate of complex images with light changes, and has obvious advantages.

关 键 词：光照图像 数据挖掘技术 双边滤波算法 双树复小波域变换 

分 类 号：TN247[电子电信—物理电子学]