局部阴影条件下基于梯度直方图的模糊相关图像检索

在局部阴影条件下对图像进行检索会受到多种因素的干扰,影响检索结果的可靠性。针对局部阴影条件,提出一种基于梯度直方图的模糊相关图像检索方法,采用ICA降噪方法对局部阴影条件下的图像进行降噪处理,依据频域对图像进行增强处理。通过Sobel算子提取图像边缘,求出水平方向与垂直方向的亮度差分相近值,确定出每个方向关键边缘像素的数量,建立图像边缘梯度直方图。依据梯度直方图,通过斯皮尔曼等级相关系数求出检索图像与图像库内模糊相关图像的相关系数,确定出图像间的相近度,将相近度最高的图像看作模糊相关图像检索结果。经实验验证,所提方法检索效果好,检索性能高。

Fuzzy correlation image retrieval based on gradient histogram under local shadow

The retrieval of images under the condition of local shadow will be affected by many factors, which will affect the reliability of the results. For partial shadow conditions, proposes a fuzzy gradient histogram image retrieval method based on ICA, using the method of image denoising under partial shadow noise reduction processing, based on the frequency domain for image enhancement. Through the Sobel operator to extract the image edge, the horizontal and vertical direction of the brightness difference between the near value, determine the direction of each edge of the number of critical pixels, the establishment of the image edge gradient histogram. According to the gradient histogram, Spielman correlation coefficient obtained by correlation coefficient of image and image database retrieval in fuzzy image, determine the image are similar, as the image will be near the highest degree of fuzzy image retrieval results. The experimental results show that the proposed method has good retrieval performance and high retrieval performance.