Song Huansheng

A New Adaptive Multistage Median Filter

A new filter structure called adaptive multistage median filter (say AMM for simple) is proposed. It follows the basic architecture of the multistage median filter (MLM), but its subfilters are based on the date-dependent and sizes variant subwindows, that allows AMM give better performance than MLM. In addition, like MLM, the proposed filter is not only computationally effective, but also needs no a priori information about image or noise. Experiment results are given to show the superiority of the new proposed filters.

The high performance car license plate recognition system and its core techniques

In this paper, a high performance car license plate recognition system and its some core techniques have been introduced briefly. The introduced significant techniques include: 1) novel lighting and imaging mechanics; 2) newly developed algorithms for license plate location; 3) deformable template based character segmentation techniques. For license plate location from real nature light image, we present MGD model of car plate. Using this model, coarse and accurate license plate location algorithm has been developed. The location rate of the newly developed algorithm has reached 97.8%. The introduced systems have been used widely in china. In real all-weather application, it has shown over 97% recognition rate. This excellent system can be used in expressway toll , car monitoring and electronic-police system.

Measurement of internal pressure of vacuum tubes by micro-discharge and emission current

The traditional methods for measuring internal pressure P of vacuum tubes of breakers are the magnetron and power frequency voltage withstand test. A magnet coil is needed by magnetron and only very bad tubes can be picked out by voltage withstand test. In this paper, a new method - inception voltage V_d of micro-discharge / inception voltage V_e of emission current, i.e. V_d/V_e method, is presented. It is found that by means of the relationship V_d/V_e~P, internal pressure P can be measured accurately although vacuum gap d vary from 0.2 mm~0.3 mm. By this new method, internal pressure of vacuum tubes can be measured without using magnetic coil, but with following process: firstly, when contacts is under closed situation, by screwing the nut to stretch compression spring, vacuum gap 0.2 mm~0.3 mm is obtained, secondly power frequency voltage is applied to the gap and gas adsorption layers are polished by micro-discharge current, finally the V_d and V_e of the vacuum tube is measured. Theoretical study show that the bigger the V_d/V_e is, the smaller the P is. So by measuring the value of V_d/V_e, the value of P can be obtained. The relationships of V_d,V_e,V_d/V_e with P were measured in laboratory, the test results show that the measuring ranges of P can be 10⁰ Pa ~10^-3 Pa by this new method.

A Vehicle Trajectory Analysis Approach Based on the Rigid Constraints of Object in 3-D Space

A reliable and effective trajectories similarity metric is one of key factors for vehicle trajectories clustering problem. A trajectory clustering algorithm based on the rigid constraints of vehicles in 3-D space is proposed in this paper, which conducts vehicle trajectories clustering effectively and precisely by using a new 3-D trajectories similarity metric. Based on two key procedures, camera calibration and a reconstruction of 2-D trajectories in 3-D space, a valuable principle that the heights of the trajectories have a linear relationship between them is found through using the kinematic properties of vehicle rigid body in moving. A more valuable information need to be pay attention is that the height of two trajectories that with displacement difference satisfies a plane surface character in 3-D space when conducts a height enumeration. The experimental results show that the trajectories are very stable and reliable for clustering and event detection when reconstructing their relative position in 3-D world coordinate system.A reliable and effective trajectories similarity metric is one of key factors for vehicle trajectories clustering problem. A trajectory clustering algorithm based on the rigid constraints of vehicles in 3-D space is proposed in this paper, which conducts vehicle trajectories clustering effectively and precisely by using a new 3-D trajectories similarity metric. Based on two key procedures, camera calibration and a reconstruction of 2-D trajectories in 3-D space, a valuable principle that the heights of the trajectories have a linear relationship between them is found through using the kinematic properties of vehicle rigid body in moving. A more valuable information need to be pay attention is that the height of two trajectories that with displacement difference satisfies a plane surface character in 3-D space when conducts a height enumeration. The experimental results show that the trajectories are very stable and reliable for clustering and event detection when reconstructing their relative position in 3-D world coordinate system.

A Point-Based Tracking Algorithm for Vehicle Trajectories in Complex Environment

In this paper, a point-based tracking algorithm is presented, which can be used in traffic jams and complex weather conditions. The main approaches for tracking vehicle trajectories are based on accurately segment for the moving vehicles, while uneven illumination, shadows and vehicle overlapping are difficult to handle. The main contribution of this paper is to propose a point tracking algorithm for vehicle trajectories without a difficult image segmentation procedure. In the proposed algorithm, feature points are extracted using an improved Moravec algorithm. A specially designed template is used to track the feature points through the image sequences. Then trajectories of feature points can be obtained, while unqualified track trajectories are removed using decision rules. The experiment results show that the algorithm is robust enough for vehicle tracking in complex weather conditions.

Mitigating the cache data pollution by using branch path tracking

“Memory Wall” is one of the most important problems which must be solved for designing high performance processors. Efficient and intelligent cache system is key component of processors memory system. We notice that cache pollution caused by the speculative execution of memory accessing instructions in predictive path may affect the cache and processor performance seriously. Based on the analyses of the impacts of speculative execution and cache data pollution to processor performance, this paper proposes a cache pollution control technique by using branch paths tracking, which is called Contra. Firstly, we construct a branch paths tracking table to follow the cache data written by memory accessing instructions in speculative branch paths. Then, processing of storage, accessing and replacement to these data are particularly controlled in order to mitigate influence of data pollution to cache system and processor performance. Simulation result indicates that, Contra technique eliminates cache pollution and upgrades performance effectively. Relative to the baseline architecture, D-Cache hit rate increases from 0.03% to 6.69%, average 1.80%. Improvement of IPC ranges from 0.01% to 6.60%, average 2.56%.

Achieving non-polluting cache accessing by using data valid tag splitting

Rapid progress of Semiconductor fabrication provides capacious space for IC designs, but unfortunately, the slow development of design ability makes it difficult to utilize the on-chip resource efficiently. At present, more than half of die area of modern microprocessor is inhabited by cache. So, how to make use of cache space smartly and efficiently, and construct high performance memory system has become one of the most important content in processor architecture design. This paper analyses of the impacts of cache data pollution and speculative execution to processor performance, and proposes a non-polluting cache accessing technique based on data tag valid-bit splitting, which is called Pease. Firstly, we splits the valid-bit in D-Cache tag into two bits, reading data valid bit (RVB) and writing data valid bit (WVB). Secondly, according the different RVB and WVB combinations, corresponding accessing strategies to D-Cache are applied. As a result, Pease technique not only preserves the pre-fetch ability of speculative execution, but also makes the cache polluting data transparent, which means that, in no empty cache line situation, consequent data can be written into D-Cache directly, but without need to perform cache replacement operation. In other word, Pease technique makes polluting data totally harmless to D-cache. Simulation result indicates that, relative to the baseline architecture, Pease technique improves IPC from 1.05% to 8.40%, averagely 4.04%, and reduces miss rate of D-Cache from 19.05% to 48.16% averagely 29.66%.