fDSST (fast Discriminative Scale Space Tracking) belongs to the correlation filter tracking algorithm, which has a high success rate and precision, and also runs at a fast speed. However, it is still a huge challenge for the tracking scene of fast motion and similar object interference. In order to improve the performance of fDSST on the challenges above, this paper proposed the fDSSTs algorithm and the fDSSTss algorithm respectively. fDSSTs increases the response scores near the object location by fusing the fhog feature and the color statistical feature, so improved the tracking performance of fDSST in the fast moving scene. fDSSTss adds a multi-feature object association module on the basis of fDSST, which distinguishes the real object and the interference object from the object feature level, thereby maintaining the tracking of the real object. The fDSSTs is tested on the OTB50 dataset, in fast-moving scenarios, the success rate of fDSST is improved by 20.5% and the precision is improved by 22.8% compared with fDSST. The fDSSTss is tested on the test sequences of similar object interference, and the result shows that fDSSTss has better anti-similar object interference ability than fDSST, while meeting the real-time requirements. The experiments show that the improvements improve the success rate and precision of fDSST in fast object moving scenes, as well as the ability to resist similar object interference.
In order to solve the difficult problem of unmanned air vehicle(UAV) target detection in visible light images under complex sky background, this paper proposes a UAV target detection method based on frequency domain transform. First, the B-channel in the image LAB space is used to extract the sky and cloud boundary images, and then the image feature channel is used to construct a quaternion function. Secondly, Fourier transform is performed on the quaternion function to extract the amplitude spectrum and phase spectrum, then the amplitude spectrum image is subjected to multi-scale decomposition using wavelet transform in the frequency domain, the amplitude spectrum image of each scale and the phase spectrum image are combined by inverse Fourier transform, and the evaluation function is used to obtain the best scale image. Finally, the best-scale image and the boundary image are normalized to make a difference to obtain the final detection result. Experimental results show that the algorithm can effectively detect UAV targets under complex cloud background.
In this paper, we mainly studied how to calculate the energy values of stars received by ground optical systems. Then according to the characteristics of the radiation spectrum of stars and the radiation spectrum of sky background light, we analyzed and selected the observation spectrum of the image-forming system. We also studied the influence of optical system design parameters, such as field of view, focal length and aperture, on detection capability. At the same time, we analyzed and calculate the limitation of the detector's dark current, readout noise and other noise factors on the ability of receiving stars. The significance of these studies is that we provide an effective theoretical basis for designing and improving ground-based photoelectric detection system for star observation during the daytime. In addition, we used digital image processing technology to process the existing observation images and improve the quality of the image. We provide several algorithms for extracting small targets in strong background. We use threshold segmentation, morphological filtering and other methods to improve the signal-to-noise ratio of the image and then improve the detection ability of the system again. According to the simulation, the target extraction accuracy can reach 1/10 pixels when the target imaging size is 4 pixels and the signal-to-noise ratio is less than 5. Improving the detection ability of photoelectric detection system, detecting more available stars and obtaining their relative position information are the important basis for star map matching and the estimation of targets’ position and gesture.
Precise location of laser spot in laser precision measurement is always an important research direction. Laser has the characteristics of good direction and small divergence, so it is widely used in aerospace, weapon systems and optical measuring and testing instruments. The accuracy of the laser spot center location can directly determine the precision of measurement. Aiming at positioning the center of laser spot, in the foundation of researching the limitation of the practical application of the common laser spot center location algorithm, this paper proposes a method of laser spot center localization based on sub-pixel interpolation, which can effectively improve the signal noise ratio (SNR) of laser spot image, reduce the influence of the background noise and thermal noise. The algorithm firstly uses the threshold value decision to exclude the interference of the light of the image, and then use the improved sub-pixel interpolation algorithm for image edge detection to obtain the edge image, and finally using the circle fitting method to obtain the positioning center. Through the experiment of processing of laser spot image, the results show that improved algorithm proposed in this paper has higher positioning accuracy than the traditional centroid, and satisfies the need of laser precision measurement in reliability, positioning accuracy and noise resistance and other aspects, at the same time, the computational complexity of this algorithm is low, can greatly save the system resources, and it can be used for the processing of the video images in the hardware and software.
The image obtained from space-based vision system has increasingly high frame frequency and resolution, and field of view is also growing. Due to the dramatic increase of data scale and the restriction of channel bandwidth between satellite and ground, on-orbit data compression becomes the core of on-satellite data processing. The paper analyzes the new generation static image compression standard JPEG2000 and the key two-dimensional (2D) discrete wavelet transform (DWT) technology. Then an FPGA (Field Programmable Gate Array)implement method for 2D integer wavelet transform is designed. It adopts the spatial combinative lifting algorithm (SCLA), which realizes the simultaneous transformation on rows and columns. On this basis, the paper realizes wavelet decomposition for images with a resolution of 6576*4384 (which is divided into 1024*1024) on the FPGA platform. In particular, the test platform is built in ISE14.7 simulation software, and the device model is xc5vfx100t. The design has passed the FPGA verification. In order to verify the correctness of the algorithm, the results are compared with that obtained by running matlab code. The experimental results show that the design is correct and the resource occupancy rate is low.
An image sharpness assessment method based on the property of Contrast Sensitivity Function (CSF) was proposed to realize the sharpness assessment of unfocused image. Firstly, image was performed the two-dimensional Discrete Fourier Transform (DFT), and intermediate frequency coefficients and high frequency coefficients are divided into two parts respectively. Secondly the four parts were performed the inverse Discrete Fourier Transform (IDFT) to obtain subimages. Thirdly, using Range Function evaluates the four sub-image sharpness value. Finally, the image sharpness is obtained through the weighted sum of the sub-image sharpness value. In order to comply with the CSF characteristics, weighting factor is setting based on the Contrast Sensitivity Function. The new algorithm and four typical evaluation algorithm: Fourier, Range , Variance and Wavelet are evaluated based on the six quantitative evaluation index, which include the width of steep part of focusing curve, the ration of sharpness, the steepness, the variance of float part of focusing curve, the factor of local extreme and the sensitivity. On the other hand, the effect of noise, and image content on algorithm is analyzed in this paper. The experiment results show that the new algorithm has better performance of sensitivity, anti-nose than the four typical evaluation algorithms. The evaluation results are consistent with human visual characteristics.
KEYWORDS: Signal processing, Motion models, Detection and tracking algorithms, Servomechanisms, Telescopes, Computer simulations, Data modeling, Data storage, Clouds, Optical tracking
A moving target should be missing from a photoelectric theodolite tracker, when the clouds and other special conditions encountered in the course of a theodolite tracking a moving object, and this condition should cause the interruption of tracking process. In view of this problem, an algorithm based on the frame of parameter identification and rolling prediction to trajectory was presented to predicting the target trajectory when it missing. Firstly, the article makes a specification of photoelectric theodolite and it operating mechanism detailed. The reasons of flying target imaging disappear from the field of theodolite telescope and the traditional solution to this problem, the least square curve fitting of trajectory quadratic function of time, were narrated secondly. The algorithm based on recursive least square with forget factor, identify the parameters of target motion using the data of position from single theodolite, then the forecasting trajectory of moving targets was presented afterwards ,in the filtering approach of past data rolling smooth with the weight of last procedure. By simulation with tracking moving targets synthetic corner from a real tracking routine of photoelectric theodolite, the algorithm was testified, and the simulation of curve fitting a quadratic function of time was compared at the last part.
Aim at one application question: only a single tracking algorithm can not achieve reliable tracking in the entire trip, the
distributional fusion system based on two kind of track pattern is proposed in this paper. Two kinds of track pattern are
respectively based on the region match related track pattern and based on the characteristic point track pattern. The
experiment result testify: the multimode tracking system in this article is availability, and obtain higher levels of
tracking performance and tracking stability.
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.