In digital image forensics, the detection of exogenous elements in an image is a difficult open problem, and several ways to solve it have already been proposed. In this paper, we use differences in the intrinsic noise in digital images to highlight this type of forgery. Our method analyses an histogram based on the noise density function at the local level in order to reveal suspicious areas, by using a new tool exploiting the uniform property of the Gaussian noise contained in JPEG images. Although the use of noise discrepancies to detect splicing has already been done multiple times, most existing methods tend to perform poorly on the current generation of high quality images, with high resolution and low noise. The improvements of our method over other state of the art approaches is demonstrated on a large set of randomly spliced natural images.