Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2022-04-07 Cooperative journals: 《计算机应用研究》
Abstract： Aiming at solving problems of dislocation and artifacts in the stitching of large parallax UAV images, this paper proposed a method to find the best stitching seam based on quality evaluation QEB-U. First, the initial seam was estimated by conventional seam estimation. Then, a new quality evaluation function was proposed for evaluating each pixel on the seam according to the characteristics of the UAV image like the structural similarity, color difference, and texture complexity. The difference cost was updated according to the evaluation result and the seam was acquired after that. The estimation and evaluation process was repeated until the seam becomes stable that the iteration was stopped. Finally, the stitching UAV image was generated by gradient fusion. The experimental results showed that the proposed method can avoid dislocation and artifacts in the stitching of large parallax UAV images, which was superior to several current UAV image stitching methods. Besides, the obtained stitching seam preferentially passed through roads, woodlands, etc. It was more in line with human visual perception and performed well in common image clarity evaluation indicators.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-06-19 Cooperative journals: 《计算机应用研究》
Abstract： For the variety of screw refrigeration compressors and the complexity of faults, it is difficult to obtain effective prediction model. A prediction model of hybrid kernel least square support vector machine (LSSVM) based on adaptive differential evolution algorithm (SADE) is proposed, which has the advantages of simple theoretical structure, less parameter setting and strong searching ability than other intelligent optimization algorithms. In the course of evolution, SADE is adaptive to the difference strategy, scaling factor and crossover probability respectively, which ensures the global searching ability and population diversity at the initial stage of optimization, improves the local searching ability and the convergence speed , and uses SADE to search the parameters of kernel parameters, LSSVM parameters and the mixed kernel regulation parameter, Improving the accuracy of the hybrid kernel LSSVM prediction model. The model is applied to the compressor fault prediction, and the experimental results show that the model can predict the compressor fault effectively and verify the feasibility of the model.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-24 Cooperative journals: 《计算机应用研究》
Abstract： Under bad weather conditions, such as heavy haze, the visual effect of obtained images is poor. This paper proposeda fast restoration method of haze images based on visual perception, and measured two important parameters of atmospheric optical physical model. Firstly, in order to fit a relatively accurate atmospheric light value, this paper used threshold segmentation combined with binary tree segmentation method, and then adopted adaptive Gaussian filtering and tone adjustment methods to optimize the transmittance. The experimental results show that In comparison with the algorithm [1, 3, 5, 8], the effectiveness of the dehazing results of our algorithm is saturated and clear. Besides, it also can retain clear edge details , high contrast, high processing efficiency and can meet the actual application requirements.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-20 Cooperative journals: 《计算机应用研究》
Abstract： For the binary phase shift keying(BPSK) signal always get high bit error rate(BER) when demodulated in a low signal-to-noise ratio(SNR) environment, a K-means clustering demodulation method based on chaos Duffing oscillator has been proposed to reduce the BER. This study according to two characteristics of Duffing oscillator, not only is it very sensitive to the initial value, but also has certain immunity to noise. The phase of BPSK signal change between 0° and π cause state change in the Duffing oscillator’s phase orbits, and using a K-means clustering method to find the centroid of phase orbit by iteration, then BPSK signal can been demodulating by calculate the distance of centroids. Simulation results show that the K-means clustering demodulation method has a great improvement in the calculation rate and the demodulation accuracy in a low SNR environment.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-20 Cooperative journals: 《计算机应用研究》
Abstract： In order to achieve greater safety in RSA encryption algorithm in the data encryption process, this paper made some improvements based on in-depth analysis on the traditional RSA algorithm. It put forward a more secure method with changing traditional double primes into four primes in the traditional RSA algorithm and eliminating the requirement for transferring the product n of two random primes in public key on the basis of parameter substitution. Instead, this paper replaced the original parameter n to a new introduced parameter x. In order to boost the efficiency in the introduced algorithm, this paper used the CRT(Chinese remainder theorem) to expedite the exponentiation of large numbers. The results demonstrates the feasibility of the improved algorithm, provides a safer path for messages transmission through public key encryption. At the same time, this paper carried out a comparative analysis of encryption and decryption time among the improved algorithm and the traditional double primes and four primes RSA algorithms. The results shows that the improved algorithm to a certain degree enhanced the encryption and decryption efficiency between the message sender and receiver.
Hits
Hits
Downloads
Comment