One of the most useful thresholding techniques using gray-level histogram an image is Otsu method. The objective this paper to extend it 2-dimensional histogram. method utilizes information each pixel and its spatial correlation within neighborhood. This was compared with 1-dimensional It found that proposed performs much better when images are corrupted by noise.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A new texture-feature description operator, called the multidirectional binary patterns (MDBP) is proposed in this paper. The operator can extract detailed distribution of textures local regions by comparing differences gray levels between neighboring pixels. Moreover, texture expression ability enhanced focusing on features linear neighborhood image multiple directions. MDBP was modified introducing a “uniform” pattern to reduce grayscale values image. Combining and gray-level co-occurrence...

The author designs and implements a real-time monitoring system for inspecting car with MapX based on the technologies of RFID, GPS GIS. In thesis, architecture function frame are introduced, then key implementation discussed in detail. practice proves that is stable efficient. could not only offer precise position but also monitor information RFID cards surrounding vehicles traffic department, which supports scientific, effective prompt decision related managers. has good prospects control,...

The wing-beat frequency is one of the most important behavioural parameters migratory insects and widely used to distinguish insect species due diversity among different species. Traditional radar has been able achieve automatic extraction frequency. However, its low-range resolution fixed threshold detection may cause false alarms reduce precision authors' Ku-band system enables high-resolution, full-polarised insects, but not achieved yet. authors propose an algorithm based on a system....

Dynamic changes of the wetland in Qinghai-Tibet Plateau were analyzed by using images TM (Thematic Mapper) 1990, ETM+ (Enhanced Thematic Mapper Plus) 2000 and CBERS (China Brazil Earth Resource Satellite) 2006 as main data source this paper. And through man-computer interactive visual interpretation different periods was obtained. The results show that existing area is about 131894.18 km2 Plateau. total decreased significantly first then increased obviously over past 16 years lake has most...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

This study introduces a novel insect speed extraction method based on high resolution and full polarisation radar with vertical-looking mode. The echo model of the is first established. Then, an iterative algorithm for estimating horizontal insects derived second-order polynomial approximation. Finally, experimental results show effectiveness feasibility this in extracting insects. Compared traditional method, here more simple effective greatly reduces computation.

The purpose of this study is to demonstrate that the fully polarimetric radar can be effectively used measure insect orientation. First, polarisation scattering matrix (PSM)-based orientation extraction method reviewed. This was proposed for traditional rotation radar, but also applicable radar. Then, both anechoic chamber and field experiments were conducted PSMs orientations, which verified effectiveness based on measurement.

This paper studies influencing factors on the structure of clouds and availability satellite remote sensing images base that, puts forward that cloud area, thickness fragment are main factors. Combined with immune coding algorithm, area can be calculated; impact patterns by atmospheric scattering, refraction, so on, a method calculating is proposed; analyzing effects shape distribution amount information available for images, computing formula using large number test data, we studied their...

The train timetable plays a crucial role in the railway system, because of its high speed, large capacity, safety, comfort, and environmental friendliness. In this work, we propose an advanced multi-agent deep reinforcement learning (MDRL) method for solving timetabling problem (TTP). An improved environment built upon previous work is proposed to model TTP as Markov decision process (MDP) incorporate new feature predict inevitable conflicts future. This generates more effective experiences...

We present a multi-agent deep reinforcement learning (MDRL) framework for tackling the train timetabling problem (TTP) in this paper. A environment is proposed to model TTP as Markov decision process (MA-MDP) and incorporate new function predict inevitable conflicts future caused by some seemingly good or feasible actions present. This innovative prevents from producing deceptive experiences, which affects agent training efficiency. use novel MDRL algorithm called transformer (MAT), combines...