Remote sensing image combines the characteristics of visible light image, infrared image and other multi-spectral images, making it rich in details and low resolution. However, due to factors such as weather and transmission errors, salt and pepper noise is prone to occur, and it is difficult to effectively detect weak edges. Aiming at the problems, a morphological edge detection algorithm based on hierarchical multi-scale is proposed in this paper. Firstly, the adaptive median filter is used to smooth the remote sensing image. Secondly, it is proposed to use the mutual information of images as the cost function for processing, and the multi-scale hierarchical ratio is 2 1 . Then, the four directional structural elements of 0 , 45 , 90 and 135 are used to extract the original-scale and small-scale image edges respectively. Calculate the sum of the gray difference values of eight neighborhoods of the edge points, thus calculate the direction adaptive weight, and then the edge detection results are obtained by fusion respectively. Finally, the pixels in the edge image are classified, and the fusion method of enhancing edge and weak edge and filtering false edge is proposed, and then the edge detection image is obtained. Aiming at the application of remote sensing images, the comparison of the results shows that the proposed algorithm has stronger anti-noise performance, the weak edge detection ability is improved, thus avoiding missed detection and false detection of edge information, and detecting more complete and accurate edge details.
The spectral absorption characteristics of hemoglobin determine that the contrast between R and B components in the white light endoscopic blood vessel image is poor, and the blood vessel features in the G component are the clearest and the contrast is good. Based on this feature, this paper proposes to use the G-component image to perform nonlinear stretching to obtain the stretched image, and subtract the original image from the stretched image to obtain the G-component high-frequency detail image containing blood vessel feature information; Then, using the high-frequency detail image to perform unsharp mask processing on the R, G, and B components of the original image, respectively, to obtain a blood vessel contrast-enhanced image; In order not to cause grayscale dispersion in the transition zone of the blood vessel edge during stretching, Performance simulation experiments are carried out for endoscopic images of fundus and oral cavity. The results show that the proposed algorithm not only improves the image contrast, but also has a better enhancement effect on small blood vessels with inconspicuous original features. By comparing with the performance of Spectra B and the method of literature [6], the average gradient value of the algorithm in this paper is increased by about 300%, the information entropy value is increased by about 30%, and the DV-BV value is increased by about 75%.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.