Search

Images are widely used in engineering. But, some images such as medical ultrasound images are mainly degraded by an intrinsic noise called speckle. Therefore, de-speckling is a main pre-processing stage for degraded images. In this paper, we suggest three phases and three denoising filters. In the first phase, the coefficient of variation is computed from the noisy image. Next, fuzzy c-means (FCM) is applied to the coefficients of variation. Applying FCM leads to the fuzzy classification of image regions. Next, the second phase is a hybrid of the three denoising filters. Fast bilateral filter (BF) for homogeneous regions, improved the adaptive wiener filters (AWFs) and wavelet filter that are applied on homogeneous, detail and edge regions, respectively. The proposed improved AWF has been developed from the AWF. In the third phase, the output image is evaluated by the fuzzy logic approach. Thus, with three phases, the proposed method has a better image detail preservation compared to some other standard methods. The experimental outcomes show that the proposed denoising algorithm is able to preserve image details and edges compared with other de-speckling methods. [ABSTRACT FROM AUTHOR]

Reflectance variability in mountainous regions caused by steep slopes can decrease the accuracy of landcover mapping. Topographic correction aims to reduce this effect, and various techniques have been proposed to conduct such correction on satellite imagery. This paper presents the initial results of five different topographic correction techniques applied to LAPAN-A3 multispectral images, namely cosine correction, improved cosine correction, Minnaert correction, modified Minnaert correction and two-stage normalization. The widely-available ALOS World 3D 30 meter DEM was employed, with the evaluation made in a mountainous area in South Sulawesi, Indonesia, located in an ancient volcanic region, with slopes ranging from 0 to 60 degrees. The slope aspect was almost equally distributed in all directions. Visual and statistical analysis was conducted before and after the topographic correction to evaluate the results. Standard deviation (SD) and the coefficient variation (CV) were calculated; the results show that the topographic corrections were able to reduce the effect of shadows and relief. Minnaert correction proved to be the best method in terms of visual appearance and spectral variability reduction. [ABSTRACT FROM AUTHOR]

To accurately evaluate the development level of ecological civilization in coastal cities, this paper proposes an evaluation algorithm of coastal city ecological civilization development based on Improved BP neural network, constructs the evaluation index system, and establishes the BP neural network-based evaluation model. By collecting the relevant data of 30 cities in China Line standardization, as the input learning matrix, entropy and coefficient of variation are used to evaluate the development level of ecological civilization. Through 5000 training times, the error accuracy is close to 10–7, and the cities with low ecological civilization level and less cultivated land resources per capita are found out. Based on the improved BP neural network method, according to the development level of coastal cities, the ecological culture of coastal cities is put forward as a whole. The countermeasures and suggestions of construction are given. • In this study, the ecological civilization development level of the coastal cities in Jiangsu was evaluated. • The evaluation results showed the overall development level of ecological civilization. • In the assessment, the comprehensive level of the secondary indexes is better than that of Donghai and Guanyun. • Jiangsu coastal cities have higher ecological civilization development levels, and Donghai and Guanyun is relatively low. [ABSTRACT FROM AUTHOR]

Abstract: In this paper, a weight selection procedure in the --means algorithm is proposed based on the statistical variation viewpoint. This approach can solve the --means algorithm’s problem that the clustering quality is greatly affected by the initial value of weight. After the statistics of data, the weights of data are designed to provide more information for the character of --means algorithm so as to improve the precision. Furthermore, the corresponding computational complexity is analyzed as well. We compare the clustering results of the --means algorithm with the different initialization methods. Results from color image segmentation illustrate that the proposed procedure produces better segmentation than the random initialization according to Liu and Yang’s (1994) evaluation function. [Copyright &y& Elsevier]