Suchergebnisse

Motion estimation, object detection, and tracking have been actively pursued by researchers in the field of real time video processing. In the present work, a new algorithm is proposed to automatically detect objects using revised local binary pattern (m-LBP) for object detection. The detected object was tracked and its location estimated using the Kalman filter, whose state covariance matrix was tuned using particle swarm optimisation (PSO). PSO, being a nature inspired algorithm, is a well proven optimization technique. This algorithm was applied to important real-world problems of partially-occluded objects in infrared videos. Algorithm validation was performed by realizing a thermal imager, and this novel algorithm was implemented in it to demonstrate that the proposed algorithm is more efficient and produces better results in motion estimation for partially-occluded objects. It is also shown that track convergence is 56% faster in the PSO-Kalman algorithm than tracking with Kalman-only filter.

Surveillance of maritime domain is absolutely vital to ensure an appropriate response against any adverse situation relating to maritime safety or security. Electro-optic search and track (EOST) system plays a vital role by providing independent search and track of potential targets in marine environment. EOST provides real-time images of objects with details, required to neutralise threats. At long range, detection and tracking capability of EOST degrades due to uncertainty in target signatures under cluttered scenario. Image quality can be improved by using suitable sensors and enhancement using the target/background signature knowledge. Robust tracking of object can be achieved by optimising the performance parameters of tracker. In the present work, improvement in the performance of EOST subsystems such as sensor, video processor and video tracker are discussed. To improve EOST performance in terms of detection and tracking, sensor selection criterion and various real time image processing techniques and their selection criteria for maritime applications have been also discussed. Resultant improvement in the quality of image recorded under marine environment has been presented.

In modern electro-optical systems, infra-red (IR) imaging system is an essential sensor used for day and night surveillance. In recent years, advancements in IR sensor technology resulted the detectors having smaller pitch, better thermal sensitivity with large format like 640.512, 1024.768 and 1280.1024. Large format IR detectors enables realisation of high resolution compact thermal imager having wide field-of view coverage. However, the performance of these infrared imaging systems gets limited by non uniformity produced by sensing element, which is temporal in nature and present in spatial domain. This non uniformity results the fixed pattern noise, which arises due to variation in gain and offset components of the each pixel of the sensor even when exposed to a uniform scene. This fixed pattern noise limits the temperature resolution capability of the IR imaging system thereby causing the degradation in system performance. Therefore, it is necessary to correct the non-uniformities in real time. In this paper, non uniformity correction algorithm and its implementation in reconfigurable architectures have been presented and results on real time data have been described.

Range performance of a thermal imaging system is characterised by the prevailing atmospheric condition present at that time. There are two dominant parameters that limit the range performance of any thermal imaging systems i.e. ambient temperature and relative humidity. In the present work, comparative study of acquisition range performance of thermal imaging system operating in LWIR and MWIR spectral bands has been presented as a function of absolute humidity (AH) which is responsible for attenuation of IR radiation due to water vapour molecules present in path length. Presentation of acquisition range as function of AH leads to a single range performance table/graph for thermal imaging system under consideration for predefined visibility (V), target size, ambient temperature (T), target to background temperature difference (ΔT) and relative humidity (RH). This table/graph can be used to predict detection, recognition and identification ranges for any set of combination of air temperature (T) and relative humidity (RH). The approach presented in this paper is versatile and has been illustrated through comparative performance analysis of LWIR and MWIR thermal imaging systems based on 640X512 staring focal plane array (FPA) having identical design parameters in terms of resolution (IFOV). It has been shown that MWIR performance is superior to LWIR beyond a crossover value of AH(T) even though MRTD of MWIR sensor is inferior to that of LWIR sensor at all spatial frequencies. Study has been carried out both for clear atmosphere and hazy conditions.