The Experts below are selected from a list of 294 Experts worldwide ranked by ideXlab platform
Yi Zhang - One of the best experts on this subject based on the ideXlab platform.
-
ICIA - Parallel implementation of Low Light Level image enhancement using CUDA
2015 IEEE International Conference on Information and Automation, 2015Co-Authors: Peiyi Shen, Yi Zhang, Liang Zhang, Juan Song, Xilu Peng, Kang YiAbstract:Enhancement algorithms can make Low Light Level images have a clear visual effect like the one captured during the daytime, but due to high complexity and generous computational cost, Low Light Level image enhancement algorithms are usually difficult to meet real-time requirements which make it difficult to be widely used in practical application. For this situation, a parallel optimization algorithm of Low Light Level image enhancement using CUDA is proposed. Enhancement algorithm based on de-hazing technique is used and on CPU-GPU heterogeneous platform the part of atmospheric Light estimation which is not suitable for parallel computing is improved to obtain high parallelism degree. By comparing the performance of the algorithm on GPU with CPU, we indicate that the algorithm proposed has a significant improvement in execution speed while maintaining the visual effect of the traditional algorithm.
-
Parallel implementation of Low Light Level image enhancement using CUDA
2015 IEEE International Conference on Information and Automation, 2015Co-Authors: Peiyi Shen, Yi Zhang, Liang Zhang, Juan Song, Xilu Peng, Kang YiAbstract:Enhancement algorithms can make Low Light Level images have a clear visual effect like the one captured during the daytime, but due to high complexity and generous computational cost, Low Light Level image enhancement algorithms are usually difficult to meet real-time requirements which make it difficult to be widely used in practical application. For this situation, a parallel optimization algorithm of Low Light Level image enhancement using CUDA is proposed. Enhancement algorithm based on de-hazing technique is used and on CPU-GPU heterogeneous platform the part of atmospheric Light estimation which is not suitable for parallel computing is improved to obtain high parallelism degree. By comparing the performance of the algorithm on GPU with CPU, we indicate that the algorithm proposed has a significant improvement in execution speed while maintaining the visual effect of the traditional algorithm.
-
ICNC - Simulation of the single channel dual-spectrum Low Light Level instrument
2010 Sixth International Conference on Natural Computation, 2010Co-Authors: Chuang Zhang, Yi ZhangAbstract:The single channel dual-spectrum Low Light Level (LLL) instrument uses stripe Light filter in the sole channel to obtain the stripe images which contain two LLL wave bands information. This paper specifies the simulation of the LLL stripe image, the method of separating and compensating the stripe image to the entire LLL short wave image and the LLL full wave image. The LLL stripe image is simulated by software from collecting information of dual-channel for the same scenery. The simulated LLL stripe image is separated and compensated by the method of inter-frame and in-frame, and the entire LLL images are rebuilt. The entire LLL short wave image and the LLL full wave image are fused to simulate the result of the single channel dual-spectrum Low Light Level (LLL) instrument. The simulation experimental results indicate that the single channel dual-spectrum color Low Light Level instrument has applied effectively, and has achieved the goal of the dual-band Low Light Level image fusion and objects enhancement.
-
Simulation of the single channel dual-spectrum Low Light Level instrument
2010 Sixth International Conference on Natural Computation, 2010Co-Authors: Chuang Zhang, Yi ZhangAbstract:The single channel dual-spectrum Low Light Level (LLL) instrument uses stripe Light filter in the sole channel to obtain the stripe images which contain two LLL wave bands information. This paper specifies the simulation of the LLL stripe image, the method of separating and compensating the stripe image to the entire LLL short wave image and the LLL full wave image. The LLL stripe image is simulated by software from collecting information of dual-channel for the same scenery. The simulated LLL stripe image is separated and compensated by the method of inter-frame and in-frame, and the entire LLL images are rebuilt. The entire LLL short wave image and the LLL full wave image are fused to simulate the result of the single channel dual-spectrum Low Light Level (LLL) instrument. The simulation experimental results indicate that the single channel dual-spectrum color Low Light Level instrument has applied effectively, and has achieved the goal of the dual-band Low Light Level image fusion and objects enhancement.
-
Research on image separation and reconstruction method of single channel double spectrum Low Light Level system
International Symposium on Photoelectronic Detection and Imaging 2009: Advances in Imaging Detectors and Applications, 2009Co-Authors: Chuang Zhang, Yi ZhangAbstract:On the base of analyzing the dual-channel systems, a single-channel dual-band false color night imaging principium based on inter-frame compensation is proposed to realize originally dual-channel dual-band on single-channel system with a raster filter setting forward. Single channel double spectrum Low Light Level system can receive stripe image included two spectrum information of Low Light Level in single channel by stripe filter slice. On the aspect of the dual-band Low Light Level images separation and compensation, the technique of spectrum separation and compensation reconstruction were researched, the stripe Low Light Level images were obtained from the scenery actual imaging with the 'long' wave information and the 'short' wave information in turn. The sample block compensation method based on the correlation of the gray space and the inter-frame compensation methods were designed to compensate the split dual-band images. The simulation experimental study of the dual-band Low Light Level image separation and the compensation technique were done, the results indicate that the above methods in the single channel dual-spectrum color Low Light Level system have applied effectively, and have achieved the goal of the dual-band Low Light Level image separation and the compensation.
Jianzhong Su - One of the best experts on this subject based on the ideXlab platform.
-
photon counting range intensity image strategy in Low Light Level environments
Optics Letters, 2014Co-Authors: Zijing Zhang, Yuannan Xu, Long Wu, Yong Zhang, Yuan Zhao, Jianzhong SuAbstract:We present a photon counting range-intensity image strategy based on a single-photon detector in Low-Light Level environments. In this Letter, a composite modulation method over the pulse sequence was used for the first time, to the best of our knowledge, which combined pulse-position modulation and pulse-intensity modulation. This composite modulation method could obtain range and intensity of the detected target at the same time. Besides, angle-angle information could be provided from the scanner or detector array. Thus, a range-intensity image of the target became feasible. For demonstrating this photon counting range-intensity image strategy, a proof-of-principle laboratory system was established. In Low-Light Level environments, a range-intensity image of multiple similar targets was obtained successfully with the range accuracy of centimeter Level and intensity error of 1%. Compared with the range image, a range-intensity image could better reorganize and identify similar targets.
Chuang Zhang - One of the best experts on this subject based on the ideXlab platform.
-
ICNC - Simulation of the single channel dual-spectrum Low Light Level instrument
2010 Sixth International Conference on Natural Computation, 2010Co-Authors: Chuang Zhang, Yi ZhangAbstract:The single channel dual-spectrum Low Light Level (LLL) instrument uses stripe Light filter in the sole channel to obtain the stripe images which contain two LLL wave bands information. This paper specifies the simulation of the LLL stripe image, the method of separating and compensating the stripe image to the entire LLL short wave image and the LLL full wave image. The LLL stripe image is simulated by software from collecting information of dual-channel for the same scenery. The simulated LLL stripe image is separated and compensated by the method of inter-frame and in-frame, and the entire LLL images are rebuilt. The entire LLL short wave image and the LLL full wave image are fused to simulate the result of the single channel dual-spectrum Low Light Level (LLL) instrument. The simulation experimental results indicate that the single channel dual-spectrum color Low Light Level instrument has applied effectively, and has achieved the goal of the dual-band Low Light Level image fusion and objects enhancement.
-
Simulation of the single channel dual-spectrum Low Light Level instrument
2010 Sixth International Conference on Natural Computation, 2010Co-Authors: Chuang Zhang, Yi ZhangAbstract:The single channel dual-spectrum Low Light Level (LLL) instrument uses stripe Light filter in the sole channel to obtain the stripe images which contain two LLL wave bands information. This paper specifies the simulation of the LLL stripe image, the method of separating and compensating the stripe image to the entire LLL short wave image and the LLL full wave image. The LLL stripe image is simulated by software from collecting information of dual-channel for the same scenery. The simulated LLL stripe image is separated and compensated by the method of inter-frame and in-frame, and the entire LLL images are rebuilt. The entire LLL short wave image and the LLL full wave image are fused to simulate the result of the single channel dual-spectrum Low Light Level (LLL) instrument. The simulation experimental results indicate that the single channel dual-spectrum color Low Light Level instrument has applied effectively, and has achieved the goal of the dual-band Low Light Level image fusion and objects enhancement.
-
Research on image separation and reconstruction method of single channel double spectrum Low Light Level system
International Symposium on Photoelectronic Detection and Imaging 2009: Advances in Imaging Detectors and Applications, 2009Co-Authors: Chuang Zhang, Yi ZhangAbstract:On the base of analyzing the dual-channel systems, a single-channel dual-band false color night imaging principium based on inter-frame compensation is proposed to realize originally dual-channel dual-band on single-channel system with a raster filter setting forward. Single channel double spectrum Low Light Level system can receive stripe image included two spectrum information of Low Light Level in single channel by stripe filter slice. On the aspect of the dual-band Low Light Level images separation and compensation, the technique of spectrum separation and compensation reconstruction were researched, the stripe Low Light Level images were obtained from the scenery actual imaging with the 'long' wave information and the 'short' wave information in turn. The sample block compensation method based on the correlation of the gray space and the inter-frame compensation methods were designed to compensate the split dual-band images. The simulation experimental study of the dual-band Low Light Level image separation and the compensation technique were done, the results indicate that the above methods in the single channel dual-spectrum color Low Light Level system have applied effectively, and have achieved the goal of the dual-band Low Light Level image separation and the compensation.
Zijing Zhang - One of the best experts on this subject based on the ideXlab platform.
-
photon counting range intensity image strategy in Low Light Level environments
Optics Letters, 2014Co-Authors: Zijing Zhang, Yuannan Xu, Long Wu, Yong Zhang, Yuan Zhao, Jianzhong SuAbstract:We present a photon counting range-intensity image strategy based on a single-photon detector in Low-Light Level environments. In this Letter, a composite modulation method over the pulse sequence was used for the first time, to the best of our knowledge, which combined pulse-position modulation and pulse-intensity modulation. This composite modulation method could obtain range and intensity of the detected target at the same time. Besides, angle-angle information could be provided from the scanner or detector array. Thus, a range-intensity image of the target became feasible. For demonstrating this photon counting range-intensity image strategy, a proof-of-principle laboratory system was established. In Low-Light Level environments, a range-intensity image of multiple similar targets was obtained successfully with the range accuracy of centimeter Level and intensity error of 1%. Compared with the range image, a range-intensity image could better reorganize and identify similar targets.
Yuan Zhao - One of the best experts on this subject based on the ideXlab platform.
-
photon counting range intensity image strategy in Low Light Level environments
Optics Letters, 2014Co-Authors: Zijing Zhang, Yuannan Xu, Long Wu, Yong Zhang, Yuan Zhao, Jianzhong SuAbstract:We present a photon counting range-intensity image strategy based on a single-photon detector in Low-Light Level environments. In this Letter, a composite modulation method over the pulse sequence was used for the first time, to the best of our knowledge, which combined pulse-position modulation and pulse-intensity modulation. This composite modulation method could obtain range and intensity of the detected target at the same time. Besides, angle-angle information could be provided from the scanner or detector array. Thus, a range-intensity image of the target became feasible. For demonstrating this photon counting range-intensity image strategy, a proof-of-principle laboratory system was established. In Low-Light Level environments, a range-intensity image of multiple similar targets was obtained successfully with the range accuracy of centimeter Level and intensity error of 1%. Compared with the range image, a range-intensity image could better reorganize and identify similar targets.
