最新刊期
卷 21 ， 期 11 ， 2016
- 摘要：Brain-like computing is the device, model or methodology which emulates, simulates or imitates the structure of the brain's neural system and its information processing mechanism, with the goal to produce the brain-like computer and brain-level intelligence. This report reviews the progress and challenges of brain-like computing in the last two decades, in the world and in China, including related research in the brain science, the neuromorphic devices, neural network chips, brain inspired computing models and applications. The future development trend is also prospected. Different from the classical artificial intelligence methodologies, including the symbolism, connectionism, behaviourism and statisticsism, brain-like computing follows the imitationism: imitate the brain at structure level (non von Neumann architecture), approximate the brain at device level (neuromorphic devices emulating the biological neurons and synapses), and surpass the brain at intelligent level (mainly by autonomous learning and training rather than manual programming). At present, there is still a big gap between the practical application of brain computing and the practical application of the industry, which provides an important research direction and opportunity for the researchers.关键词：brain-like computing;neuromorphic computing;brain-inspired intelligence;neuromorphic device;neural network chip;spiking neural network6900|600|9更新时间：2024-05-07
- 摘要：The continuously improving resolution of spatial, temporal, spectral, and radiometric for remote sensing data also increases the data type. For example, the amount of remote sensing data acquired from the aerospace, aviation, space, and other remote sensing platforms is increasing dramatically. Therefore, remote sensing data have obvious big data characteristics. This study analyzes the key technologies and issues in applying remote sensing big data and provides valuable reference for researchers. Based on numerous references, the characteristics of remote sensing big data are clarified. The processing systems for remote sensing big data are introduced from the perspectives of GPU hardware acceleration, clustering, grid, cloud computing, cloud grid, and complex high performance. The key technologies of remote sensing big data, including distributed clustered storage technology, are discussed. The existing problems are discussed through uncertainties, information fusion, machine learning, and analysis platform of remote sensing big data. The development trends are also discussed, including modeling a variety of uncertainties of remote sensing big data and machine learning methods for remote sensing big data. This study reviews the characteristics of remote sensing big data, the typical processing system, and the core technology. Key issues and future trends in this area in the practical application of academic research are also summarized. Big data technologies bring not only opportunities but also challenges for remote sensing data mining and knowledge acquisition. Several significant breakthroughs, such as machine learning for big data, unified analysis framework, and high-level information fusion, can promote further development for remote sensing knowledge mining.关键词：uncertainties modeling;multi-source information fusion;machine learning;high performance computing;remote sensing big data12063|659|26更新时间：2024-05-07
- 摘要：Several tiny particulates are suspended in air during poor weather condition (e.g., haze or fog). The color and contrast of the captured picture from image devices are severely degraded because of scattering, thereby consequently affecting visual experience. Haze is a common phenomenon in China's cities and towns, especially in the metropolis. Haze reduces visibility and seriously affects the closed circuit television surveillance system, thereby leading to difficulties in traffic monitoring and increase in traffic accidents. Using traditional methods to remove the haze in traffic image usually results in various problems, such as halo and color distortion. To remove the large sky area in traffic image, this paper proposes a haze-removal method based on sky segmentation and dark channel prior. We use the maximum of the top 0.1 percent brightest pixels in the dark channel (as selected by He) corresponding to the original image as the value to the atmospheric extinction coefficient . Accordingly, the value of in each channel can be made closer to the maximum pixel value of 255. Resulting image after haze removal can generate color cast or a large number of color spots. According to the characteristics of haze image in traffic scene, we propose a novel algorithm that automatically separates sky regions to optimize the model and thus solve the image distortion of sky region after dehazing by dark channel prior. For sky segmentation, we introduce an OTSU method to complete the task. In our haze removal algorithm, we present the average intensity of the sky as the atmospheric extinction coefficient, and estimate the scene transmission of the sky and the non-sky regions. Then, we combine two parts of the scene transmission as a whole to refine the transmission. This step makes the restored sky region highly natural looking. Compared with Fattal's method and He's method, our algorithm obtains better image sharpness and edge details of the recovered image. The distortion of the recovered image is also lower than that of the recovered image by Fattal's method and He's method. In the rehabilitation of the sky region, our algorithm exhibits highly natural looking and smooth resulting image. On the contrary, Fattal's method and He's method show a large number of spots or halos appearing in the sky region. The proposed method can effectively restore the traffic haze image. Specifically, no spots or distortions are found in the sky region of the recovered image by our method. The proposed method can also provide an effective theoretical basis and technical support for road traffic supervision. However, our algorithm has the following limitations. The effect of the recovered image is poor when distant objects are under a thick-haze environment. This result is due to the optimization of scene transmittance. The attenuation coefficient of the transmittance (scattering coefficient) is assumed to be constant. However, in the actual atmosphere, the scattering coefficient changes under different weather conditions. Therefore, our future research will incorporate the weather factor into our algorithm to optimize an accurate and robust approach for removing the haze from traffic scene image.关键词：haze traffic scene;dark channel prior;haze-removal;image segmentation;image restoration;haze image formation4775|520|10更新时间：2024-05-07
- 摘要：Real-world images are often distorted by unknown noise and intensity inhomogeneity, thereby making segmentation a challenging task. The local correntropy-based K-means (LCK) model shows significant improvements on images with unknown noise and uneven gray distribution. However, the segmentation results are sensitive to the initial contour, and the speed of the segmentation convergence is slow. To solve these problems, this paper presents a new two-stage segmentation model based on LCK model. The new model is a combination of two stages, and each stage is based on LCK model. In the first step, the convolution result of image information and Gauss kernel was down-sampled, and the down-sampled result was segmented based on LCK model resulting on coarse segmentation results and coarse level set functions accordingly. The down-sampling of the original image resulted in a coarse scale image, which could reduce data size. With the benefit of data size reduction, the down-sampled image could be rapidly segmented to an approximate result. Compared with direct down-sampling operation, down-sampling with convolution of the image information and Gauss kernel lost lesser information and could calculate local weighted average. Therefore, the gray image value was suitable. In the second step, with the smoothness constraints of level set functions, the coarse segmentation results and according coarse level set functions were up-sampled to the original image scale. The coarse level set functions of up-sampling were then used as initial value of explicit segmentation based on the LCK model. Given that the initial value was a close approximate of the object boundary, less iteration was needed to obtain results. The proposed model could provide improved contour, which was close to the object boundary for LCK model. The results of segmentation of synthetic image show that, compared with LCK model, the proposed model converged faster and was more accurate. By utilizing F-score value as an evaluation criterion, the proposed model obtained higher values than the LCK model. In addition, when images were intensity inhomogeneous or distorted by different noises, the proposed model could secure improved results with iterations of less than 50, whereas iterates of the LCK model could reach at least 1000. The proposed model was more robust than the LCK model on natural and synthetic images with complex noises. A fast and accurate segmentation based on LCK model is proposed. Based on the feature of down-sampling, the processing time is reduced without losing much information. The proposed model combines down-sampling with Gauss kernel to reserve much image information. To avoid the sensitivity of LCK model to the initial contour, the coarse segmentation provides an initial contour close to real object boundary. The proposed algorithm can rapidly segment an image with unknown complex noise.关键词：image segmentation;active contours;level set method;coarse segmentation;accurate segmentation3291|378|0更新时间：2024-05-07
- 摘要：Although sparse representation-based tracking approaches show good performance, they usually fail to observe the object motion because of noise, rotation, partial occlusion, motion blur, and illumination or pose variation. This study proposes an algorithm based on sparse representation and a priori probability of object template to improve tracking capability under partial occlusion, rotation, pose change, and motion blur conditions.An L1 tracker is also developed, which runs in real time and possesses better robustness than other L1 trackers. The importance of the target template is measured by a priori probability and is considered in the proposed algorithm when updating the object template. Combined with the regularization model, a novel sparse representation model of the object is presented. Based on the proposed target appearance model, an effective template update scheme is designed by adjusting the weighs of the target templates. The tracking particles of the current frame are generated by the last tracking result according to the Gaussian distribution. The sparse representation of each particle to the template subspace is obtained by solving the L1-regularized least square problem, and a target searching strategy is employed to find the particle that well matches the template as the tracking result. The particle filter is then used to propagate sample distribution in the next tracking frame. Compared with existing popular tracking algorithms, the proposed algorithm can achieve better tracking performance in diverse test video datasets.Experimental results demonstrate that the proposed algorithm can handle appearance changes, such as pose variation, rotation, illumination,motion blur, and occlusion. Compared with state-of-the-art methods, the proposed algorithm performs well and obtains the best results in the sequences of FaceOcc1, Girl, BlurBody, and Singer1, with average center location errors of 6.8, 4.0, 16.3, and 3.5 pixels, respectively. The average tracking success rate of the proposed algorithm is high. The tracking accuracy is improved with the proposed minimization model for finding the sparse representation of the target, and the real-time performance is achieved by a new APG-based numerical solver for the resulting L1 norm-related minimization problems. The proposed algorithm can track target robustly and reliably under partial occlusion, rotation, pose variation, and motion blur conditions.A very fast numerical solver based on the accelerated proximal gradient approach is developed to solve the resulting L1 norm-related minimization problem. Qualitative and quantitative evaluations demonstrate that the performance of the proposed algorithm is comparable to that of the state-of-the-art tracker on challenging benchmark video sequences. The proposed method can therefore be used for engineering applications.关键词：target tracking;sparse representation;priori probability;particle filter;template update;regularization model2877|420|2更新时间：2024-05-07
- 摘要：Person re-identification is a very challenging problem and has practical application value. It plays an important role in video surveillance systems because it can reduce human efforts in searching for a target from a large number of videos. However, the pedestrian's image is easily affected by illumination changes, different viewpoints, varying poses, complicated background and the problem of occlusion and scale. It is likely to form a lot of differences in appearance and that causes interference in person re-identification. To solve this problem, many studies concentrate on designing a feature representation or metric learning method. For the above problem, this study proposes a robustness algorithm based on multi-features fusion and independent metric learning for person re-identification. First, the original images are processed by image enhancement algorithm to reduce the impact of illumination changes. This enhancement algorithm is committed to making the image closer to the human visual characteristics. Then, using the method of non-uniform segmentation processes images. At the same time, processed images are extracted from four color features including HSV, RGS, LAB and YCbCr feature and a texture feature of SILTP (scale invariant local ternary pattern).What's more, through multi-features fusion and independent metric learning, the algorithm gets a similarity measure function of the related person. Finally, the algorithm weights the original similarity and gets the ultimate similarity achieving person re-identification. The proposed method is demonstrated on three public benchmark datasets including VIPeR, iLIDS and CHUK01. Each dataset has its own different characteristics. And experimental results show that the proposed method achieves a higher accuracy rate with excellent features and particular method of fusion and learning compared with other similar algorithms. The proposed method achieves a 42.7% rank-1 (represents the correct matched pair) on VIPeR benchmark and respectively 43.6% and 43.7% on iLIDS and CHUK01 benchmark. It is worth mentioning that the rank-5 (represents the expectation of the matches at rank 5) are more than 70% on the three datasets. It greatly improves the recognition rate and has practical application value. The experimental results show that the proposed method can more effectively express pedestrian's image information. Furthermore, the proposed method has strong robustness to illumination changes, different viewpoints, varying poses, complicated background and the problem of occlusion and scale and effectively improves the recognition rate.关键词：person re-identification;multi-features fusion;metric learning;SILTP feature;non-uniform segmentation3813|386|4更新时间：2024-05-07
- 摘要：To reduce the effect of partial occlusion in facial expression recognition, this paper proposes a new method of facial expression recognition based on local feature fusion. First, the normalized images are processed by the Gaussian filter to reduce the effect of noise. According to their different contributions in facial expression recognition, all the images are then divided into two main parts: near the eye and near the mouth. To analyze considerable structure detail, these two parts are further divided into several non-overlapping blocks. The following two patterns are used to extract the features of each sub block: the difference center-symmetric local binary pattern, which is the change of center-symmetric local binary pattern; and the gradient center-symmetric local directional pattern, which is the change of difference local directional pattern. The features are marked as two binary sequences, which are then cascaded to obtain the characteristic histogram of the sub block. The final histogram of the image is obtained by cascading the histogram of each sub block. Finally, the nearest neighbor method is used for classification. Chi-square distance is used to calculate the distance among the characteristic histograms of the testing and training images. Considering the difference of the amount of information contained in each sub block and to reduce the effect of occlusion further, information entropy is used to weigh chi-square distance adaptively. Three cross experiments are conducted on JAFFE and CK databases. The average recognition accuracies in random occlusion, mouth occlusion, and eye occlusion cases are 92.86%, 94.76%, and 86.19% on JAFFE database, and are 99%, 98.67%, and 99% on CK database. In the aspect of feature extraction, our method describes the image from two aspects: one is the difference of the pixel values in the gradient direction, and the other is the difference of the edge response values between gradient directions. Accordingly, the image can be fully described. In the aspect of occlusion, image segmentation and information entropy are used to weigh chi-square distance adaptively. Thus, our method can effectively reduce the effect of occlusion. Under the same experimental conditions, experimental results show the effectiveness and superiority of the proposed method to other classical local feature extraction and occlusion handling methods.关键词：facial expression recognition;partial occlusion;difference center-symmetric local binary pattern (DCS-LBP);gradient center-symmetric local directional pattern (GCS-LDP);adaptively weighted4471|825|0更新时间：2024-05-07
- 摘要：Visual tracking estimates the states of a moving target in a video. This technology is the most important and fundamental topic in computer vision and has several applications, such as surveillance, vehicle tracking, robotics, and human-computer interaction. L1 object-tracking based on sparse representation expresses each target candidate as a linear combination of dictionary templates. In such tracking, the global information is considered without analyzing the local information. To overcome drifting problems in background clutter, this paper proposes a tracking method based on positive patch voting. Given the over completeness of sparse representation dictionary and sensitivity to changing local features, we present the target by a set of image patch particles to consider the local structure of target templates. Extracting image patches is the core of our algorithm and directly affects the result of tracking. Specifically, we present a tracking reliability metric to measure how positively a patch can be tracked. Accordingly, a probability model is proposed to estimate the distribution of positive patches under a sequential Monte Carlo framework. To estimate how likely a patch can be preliminarily obtained, we adopt the peak-to-sidelobe ratio as a confidence metric. This ratio is widely used in signal processing to measure the signal peak strength in response map. The confidence function is proportional to the response map of image patches, and is a distance function between the templates and patches. Instead of using computationally intensive unsupervised clustering methods to group the image patches, we simply divide the image into two regions by a rectangle box that is obtained by the L1 method centering at the target. We then formulate a similarity function to measure the patches that are inside the bounding box and a confidence score that is higher than zero. We label the patches that obtained the high score. Finally, we calculate the weight of all positive patches and vote the optimal location of the tracked target. The traditional target tracking based on sparse representation simply considers global information without analyzing the local information, and L1 object tracking easily produces drifting problems under complex situations. Thus, we present the target by a set of image patches and formulate a new patch reliability metric to extract the positive patches. Qualitative and quantitative evaluations on challenging benchmark image sequences demonstrate that the proposed algorithm can handle highly diverse and challenging situations of visual tracking. Sparse representation is applied to visual tracker by modeling the target appearance using a sparse approximation over a template set. However, the proposed method cannot adapt to complex and dynamic scenes because of various factors, such as background clutter and illumination change. By formulating the confidence and similarity function, we extract the positive patches. We finally find the target location according to voting of positive patch weight. Unlike other classical methods, our method can deal with occlusion, illumination change, and fast movement to accomplish robust tracking in sequence with background clutters.关键词：positive patch;voting;sparse representation;object tracking;particle filter2842|337|0更新时间：2024-05-07
- 摘要：The continuous intensification of diversification and polytrope on market requirement has created demand for high flexibility and adaptability of production assembly. Human-computer interaction cooperation in mechanical assembly is an effective way to satisfy the market requirement. The organic texture character of hand skin image in mechanical assembly must be detected effectively. Such detection can in turn determine the capability of the robot to recognize and comprehend the action and behavior of the operator in human-computer interaction cooperation in mechanical assembly. A recognition algorithm is presented for hand skin recognition in image based on multi-color space information. The proposed algorithm is used in human-computer interaction coordination in mechanical assembly. A multi-component mixture model of skin color in the RGB color space is established for clustering by the off-line learning method of the Gaussian mixture model and expectation maximization algorithm. The hand skin information is expressed by mixture clustering. Furthermore, the law of skin color distribution in YCrCb color space is studied by on-line fast learning to sparse Gaussian model. The hand skin recognizer is designed based on muti-color space information. Image illuminance significantly influences the likelihood value of skin region model. The high threshold value is fitted for high image illuminance region, and low threshold value is fitted for low image illuminance region. Experimental results for hand skin recognition based on on-line sparse fast learning show that the redundance of likelihood in YCrCb three-color space channels is acceptable. The Gaussian mixture model and the corresponding learning method are valid. The defect of initial recognition can be revised by the YCrCb skin sparse Gaussian mixture model. Moreover, the on-line fast learning algorithm can enhance the adaptability of the skin model to the illumination distortion and improve the results of recognition in the fixed scene. The skin region is detected based on the results of different models in multi-color space. The time complexity of on-line learning algorithm is related to the integrity of initial skin region recognition, and the illumination equilibrium degree of scene is related to the parameter update quantity of recognition model. Both factors influence the space complexity of the recognition algorithm. Hand skin color modeling and region detection based on Gaussian model has good adaptability and practicability in mechnical assembly experiment. Hand skin detection based on the proposed algorithm obtains better recognition integrity than that based on ellipse clustering in YCrCb space. The convergence time of the recognition must be short to meet the real-time constraint of recognition.关键词：hand skin detection;Gaussian mixture model;expectation maximization algorithm;online sparse learning;recognition in muti-color space3230|565|2更新时间：2024-05-07
- 摘要：Corner detection is a major research area in computer vision. Many algorithms of corner detection have been proposed in recent years. The existing methods can be divided into intensity-, contour-, and template-based approaches. The detection performance of contour-based corner detection is relatively stable, but this approach is sensitive to local variation and noise on the curve. Many contour-based image corner detectors provide only an empirical threshold to extract the corner. Thus, this paper proposes a novel contour-based corner detection with adaptive threshold, and this algorithm has a robust performance in local variation and noise on the curve. The proposed method is based on anisotropic Gaussian directional derivative (ANDD) representations, and searches the invariant property of geometry and gray level in edges and corners. The adaptive threshold of the difference between the edge and corner points is obtained by the regularization calculation. This corner detection finds the edge map using the Canny edge detector and extracts edges from the edge map first. The ANDD filter is then used to filter the extracted edge curves, and the response of each pixel is calculated and compared with the adaptive threshold value. Then, the point value that is higher than the threshold value is chosen as the candidate corner. Finally, the non-maximum suppression is applied to the candidate corner set and the final corners are obtained. The proposed detector is compared with three different detectors under affine transforms and Gaussian noise degradation. The evaluation criteria of performance are average repeatability and localization error. In simulation experiments, the average rankings of four algorithms are as follows: Harris (3.375), He and Yung (2.625), ANDDs (2.625), and the proposed method (1.375). The corner matching performance of the corner detection algorithms are compared under the noise-free and noisy environments. Experimental results show that the proposed method attains excellent performance on average repeatability and localization error under affine transforms and Gaussian noise degradation. The number of false and missed corners is less than that of the three other corner detectors in matching experiments. The proposed corner detection is a contour-based method with an adaptive threshold. Similar to most contour-based algorithms, the proposed method detects the edge map of input image using the edge detector first and then extracts the edges from the edge map. However, the proposed algorithm is different from the traditional contour-based corner detection using only the edge information. The proposed algorithm also utilizes intensity variations of pixels on edges. Furthermore, the new method uses a global adaptive threshold to avoid erroneous judgment of corners. The sum of normalized intensity variations effectively reduces the influence of noise or illumination on detection performance. As obtained from experiments of corner matching, affine transforms, and Gaussian noise degradation, the proposed method shows excellent performance in terms of detection accuracy, average repeatability, localization error, and noise robustness.关键词：corner detection;contour-based;anisotropic Gaussian directional derivatives;adaptive threshold3809|428|7更新时间：2024-05-07
- 摘要：Considering that achieving non-contact and high precision in fatigue driving detection is difficult, the method based on computer vision arises as a possible solution. The key point to achieve fatigue driving detection is to recognize the open and closed state of driver's eyes and estimate if the driver is fatigued and sleepy. This paper proposes a new fatigue driving detection method based on sclera segmentation to solve the problem of recognizing the open and closed state of driver's eyes in fatigue driving. The face is tested from the acquired picture using a face detection method based on Adaboost raised by Viola and Jones. Then, a Gauss sclera model based on YCbCr color space is built for the next step because of the good clustering performance of sclera in Cb-Cr color space. The sclera is detected in the face region using this model and the area of the sclera is calculated. The sclera area is considered the eye opening index and the fatigue state is determined along with the PERCLOS criterion. Experimental results for 10 short testing videos show that the proposed algorithm can select the sclera part and recognize the open and closed state of human eyes effectively. The accuracy rate can reach up to 96.77%. The proposed method can achieve satisfactory segmentation effect under good lighting condition. Using the sclera area as the eye opening index, the proposed method can also judge the state of driver's fatigue accurately. The new approach is effective and must be further studied.关键词：fatigue driving;face detection;sclera characteristics;YCbCr;Gauss model;percentage of eyelid closure over the pupil over time (PERCLOS)3335|357|4更新时间：2024-05-07
- 摘要：Using planar stitching algorithm causes serious distortions in the panorama image, thereby resulting in difficulty ensuring good visual consistency. Meanwhile, the traditional cylindrical stitching algorithm cannot meet real-time requirements. To overcome these shortcomings, this paper proposes a cylindrical panorama stitching algorithm based on the improved SIFT (scale-invariant feature transform) feature descriptor. First, the image sequences to be stitched are transformed using cylindrical projection, and then the improved SIFT feature detector is adopted to extract the feature points. Accordingly, 64 dimensional SIFT feature descriptors are generated. Based on the Euclidean distance of feature descriptors, the initial feature points are determined. By using RANSAC (random sample consensus) method, false matching feature points can be eliminated further and the space transformation matrix between the images to be stitched can be constructed. Finally, the image registration can be completed successfully according to the above-mentioned space transformation matrix, where the weighted average fusion method is used to realize the seamless splicing of images. This paper presents a new cylindrical image-stitching algorithm, which can effectively avoid distortion problems in planar stitching algorithm. The proposed algorithm achieves good visual consistency of panoramic image. The speed of the proposed algorithm is nearly two times of that of the traditional cylindrical image-stitching algorithm. As obtained from the experiment on panorama images using image sequences with different numbers and sizes, the proposed algorithm can stitch images more quickly and more efficiently than the planar stitching algorithm and the traditional cylindrical stitching algorithm. The generated panorama image has wide vision and high resolution, and is suitable for image-stitching applications of real-time requirements.关键词：cylindrical panorama;cylindrical projection;SIFT detectors;image registration;image fusion4125|453|4更新时间：2024-05-07
- 摘要：Indirect calibration can build a calibration model and has been used by many academic researchers. For most methods, the detection field of binocular stereo vision system has been regarded as a cuboid space to make the sample set. However, the main problem existing in the method of sample is that the detection field of binocular stereo vision system is conical space. Predicting the label of instance, which is absent in sampling area, is also difficult. The calibration model is usually not a clear mathematical model in these studies. This paper proposes a calibration method by support vector machine (SVM) based on binocular system full vision field sampling to improve the video camera's insufficient and indirect calibration sampling and vague model expression, and realize the complete modal sampling of detection view field in small vision field. The sampling method uses the readability of the target number of hexagonal lattice calibration board as the basis, collects parallax coordinates and world coordinates, and establishes complete sample sets in the entire effective vision field of the binocular system. The SVM is selected to train the sample set, and a calibration with mathematical expression is established with model parameters calculated from the SVM algorithm to the decision function. Five mutually non-centrosymmetric polygons are placed at the four corners and center of the hexagonal lattice calibration board. The position information of the calibration board can be obtained with the collected partial region information of the calibration board, and the target number is known. The target number in images is acquired by HALCON operators by moving the calibration plate, and a complete set of samples of detection field is established based on binocular system. Finally, the SVM algorithm is used to train the samples to obtain calibration models that can articulate the mathematical form of a calibration. The samples used in this study include 568 targets from all 2 839 samples, and contain ground truths in the form of accurate world coordinates for detected targets. Two kinds of sampling method are adopted in the experiment, namely, traditional sampling method for cuboid space and full vision field sampling method for conical space. Experimental results show that calibrated error was reduced by 24.51% as compared with that of the model established based on traditional sampling method. Thus, the feasibility of the proposed method was supported given that the calibration error of calibration model was reduced in the unsampled region with conventional method. A calibration method by SVM based on binocular system full vision field sampling is proposed with non-centrosymmetric polygons to identify the target number on calibration board. The proposed method improves the indirect calibration precision. The method is also feasible, robust, and suitable to indirect calibration based on binocular system in small vision field.关键词：binocular system;full vision field sampling;SVM (support vector machina);indirect calibration;calibration model;hexagonal lattice calibration board2692|353|0更新时间：2024-05-07
- 摘要：Motion generation methods based on motion controllers are widely investigated and cause difficulty in computer animation. These methods are limited by poor adaptabilities of bone parameters and style variations. Given these challenges, current motion controllers cannot rapidly adapt to various user demands. This paper proposes a multi-skeleton, multi-style-oriented locomotion controller, and describes the method for its generation. First, we use an improved proportional derivative controller for preprocessing, with the aid of heuristic method to adopt the high proportional and derivative gains during our simulation process. We then apply specific rules based on skeletal variations to tune the proportional and derivative parameters of all joints. Next, we apply the twiddle iteration algorithm to optimize the proportional and derivative parameters of hip joint for the purpose of improving motion stability. Finally, we set up several objective functions aimed at maintaining motion stability and expanding motion style diversity to optimize the target pose of the controllers, and we choose the covariance matrix adaption evolution strategy to process the optimization. Experimental results show that the proposed method can generate a series of walking controllers mapped to different skeletal parameters and styles. Furthermore, the method is efficient, stable, robust, and diverse in styles. Specifically, the stability of the method is an order of magnitude higher than that of other methods. The proposed method can generate motions with good stability and changeable styles in certain degrees. The proportional derivative parameters and target poses are generated automatically through optimization, and only initial configuration sets are required. Our method can easily be learned by users with less professional skills. This study thus improves adaptability of motion generation methods based on motion controllers, and expands the application scope of motion controllers.关键词：physical animation;locomotion controller;adaptability;stability;optimization3113|1123|1更新时间：2024-05-07
- 摘要：The validation of surface information in remote sensing image determines the performance of information extraction and quantitative applications. However, image quality evaluation based on pixel gray and human visual system cannot evaluate the expressive capability of surface information. Therefore, this study evaluates multi-spectral image quality based on the precision of reflectance and normalized difference vegetation index(NDVI) from the perspective of land surface validation. A method of evaluating multi-spectral images based on the validation of land surface parameters was designed, and synchronous observation was proposed to improve the accuracy of image quality evaluation. At the same time as the remote sensing satellite imagery, the ground measurement of atmospheric optical properties and land surface spectrum were implemented. The synchronous imaging on the experimental area of GF-1 and SPOT-7 was successfully completed, and the atmospheric optical properties and land surface spectrum were synchronously measured on the ground. The radiometric error of multi-spectral images was processed with the atmospheric optical properties to eliminate the influence of atmosphere. The equivalent reflectance and NDVI of multiple surface features were calculated using the measured spectrum. Meanwhile, the reflectance and NDVI of image were calculated through multi-spectral images using window analysis. The analysis of validation of land surface parameters was based on the statistics of absolute and relative errors in reflectance and NDVI between ground measurement and image. The multi-spectral image quality was evaluated using the error of each spectral band combined with the application in information expression. In the artificial targets, the reflectance error in four bands of GF-1 is within 5% and is more accurate than that in those of SPOT-7. In the vegetation features, the reflectance error in the blue, green, and red bands of SPOT-7 is within 4%. In the near-infrared band, the reflectance error is within 15%, the error of NDVI is within 16%, and the reflectance and NDVI of SPOT-7 are more accurate than those of GF-1. In the ground features, the reflectance error in four bands of GF-1 is within 6% and is more accurate than that in those of SPOT-7. Therefore, SPOT-7 multi-spectral image has high spectrum validation in vegetation features, and GF-1 multi-spectral image has high spectrum validation in ground features. Comparing the reflectance between SPOT-7 and GF-1 multi-spectral images shows that the recognition of vegetation features is stronger and the growth monitoring of vegetation features is weaker in SPOT-7 than in GF-1. The detection in type, water content, and surface roughness of hard features is more sensitive in GF-1 than in SPOT-7. Comparing the NDVI between SPOT-7 and GF-1 multi-spectral images shows that the performance in growth state, coverage density, and distribution of vegetation is more accurate in SPOT-7 than in GF-1. The proposed evaluation method for multi-spectral image considers the validation of land surface parameters, and can evaluate image quality effectively based on the precision in spectral information expression.关键词：image quality evaluation;synchronous observation;radiometric correction;reflectivity;normalized difference vegetation index (NDVI)3518|721|2更新时间：2024-05-07
- Postal code：100190
- Tel：010-58887035/58887030/58887418 Email：jig@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备05080539号-4
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0