阴影检测技术在裁割机样片轮廓提取中的应用研究

Study on the shadow detection applied in the sample contour extraction of cutting machine

阴影将导致轮廓提取时样片的轮廓发生扭曲变形以及部分轮廓的丢失,为了提高轮廓提取的准确性,提出了一种基于色度畸变和均值的局部多层次差分算子的阴影检测方法.新方法从彩色不变性和纹理相似性两个方面出发,通过阴影覆盖前后色度变化相对较小和纹理具有相似性的原理来确定阴影区域.实验结果表明:该方法能够有效去除样片周围的阴影,具有良好的性能.

The shadow will lead to the sample distorted and lost when the sample contour is extracted.In order to improve the accuracy of contour extraction,a shadow detection method with local multi-level difference operator based on color distortion and means is proposed in this paper.This paper analyzes the color invariance and texture similarity,then propose a shadow detection method based on chrominance distortion and local multi-level difference operator.This shadow detection method based on chrominance distortion will determine the shadow area through comparing the area with small color change before and after shadow covering and texture similarity principle.The experimental results show that this method can effectively remove the shadow around the sample and the method has a good performance.