Way Soong Lim

An Effective Surveillance System Using Thermal Camera

Thermography, or thermal visualization is a type of infrared visualization. Thermographic cameras are used in many heavy factories like metal recycling factories, wafer production factories and etc for monitoring the temperature conditions of the machines. Besides, thermographic camera can be used to detect trespassers in environment with poor lighting condition, whereby, the conventional digital cameras are less applicable in. In this paper, we proposed two simple and fast detection algorithms into a cost effective thermal imaging surveillance system. This surveillance system not only used in monitoring the functioning of different machinery and electrical equipments in a factory site, it can also used for detecting the trespassers in poor lighting condition. Experimental results show that the proposed surveillance system achieves high accuracy in monitoring machines conditions and detecting trespassers.

Home Alone Faint Detection Surveillance System Using Thermal Camera

Fainting is the unwanted event occur amongst the senior citizens, patients or pregnant women. This event can cause physical injuries or even mental problems. In this paper, we proposed a simple faint detection algorithm apply into a thermal imaging camera to monitor the fainting event. In addition, this system is to allow the immediate treatment can be carried out for life saving purposes when there are people fainting. Experimental results on sample images show that the proposed surveillance system achieved high accuracy of 96.15% for poor lighting condition and 86.19% in indoor in detecting the human faint event.

An Effective Trespasser Detection System Using Thermal Camera

This paper presents a novel human detection algorithm embedded into a thermal imaging system to form a fast and effective trespasser detection system for smart surveillance purpose. To identify a human object, pattern recognization technique is employed where human head is detected and analyzed for its shape, dimension and position. The experimental result shows that the proposed trespasser detection algorithm is fast and accurate with accuracy of 98.62% in its optimum setting with short routine time of 112 milliseconds.

Wireless Webcam Based Omnidirectional Health Care Surveillance System

This paper proposes an efficient omnidirectional surveillance system for health care purposes. The proposed surveillance system is used to detect fainted person at health care centers such as old folks home, hospital, and handicapped center and is located at the place with good light condition. When a fainted person is detected, it send immediate signal and thus alerts/treatments can be carried out. In this system, a wireless webcam attached with a hyperbolic mirror is used to capture the omnidirectional scene at health care centers. The captured scenes are then fed into a laptop computer for image processing, faint detection and alarm purposing. Log-polar mapping is used to map the omnidirectional images into panoramic images. An effective faint detection algorithm is designed to detect a fainted human at health care centers. Experimental results show that the proposed wireless omnidirectional surveillance system achieves high accuracy (92.99 % out of 10000 testing images under good light condition) in monitoring a selected private nursing home. The observed significances of this new proposed surveillance system include: omnidirectional coverage, high data compression output images, wireless, and with effective faint detection capability.

Machine condition monitoring using omnidirectional thermal imaging system

This paper proposed an efficient omnidirectional thermal imaging system for machine condition monitoring. In this thermal imaging system, the omnidirectional scenes in a machine room, production plant, pump house, laboratory, etc within a factory site are first captured using a thermal camera attached to a custom made hyperbolic IR (infrared radiation) reflected mirror. The captured scenes with some machines to be monitored are then fed into a laptop computer for image processing and alarm purposes. Log-polar mapping is proposed to map the captured omnidirectional thermal image into panoramic image, hence providing the observer or image processing tools a complete wide angle of view. A simple and effective machine condition monitoring algorithm is also designed for monitored the normal/overheat condition for machines. The observed significances of this new proposed omnidirectional thermal imaging system include: it can cover a wide angle of view (360° omnidirectional), using minimum hardware, low cost and the output thermal images are with higher data compression. Experimental results show that the proposed thermal imaging system achieves high accuracy in monitoring machine condition.

Off-time swimming pool surveillance using thermal imaging system

This paper introduces an effective intruder detection algorithm for an off-time swimming pool surveillance using thermal imaging system. Two sub-algorithms are developed for intruder detection outside and inside swimming pool regions during off-time. The algorithm can detect not only moving human in different directions, but also any water activity in the swimming pool caused by human intruder such as swimming, water-splashing and others. The dimension of human head shape is used for detecting human head while the difference of black colour pixel between successive thermal images is the criteria employed to detect any water activity due to humans. Experimental results show that the intruder detection algorithm achieved high accuracy of 95.58% for region outside swimming pool and 92.44% for region inside swimming pool.

A study of different unwarping methods for omnidirectional imaging

This paper evaluates the performance of the 3 universal un-warping methods currently applied actively around the world in transforming omnidirectional image to panoramic image, namely the pano-mapping table method, discrete geometry method (DGT) and the log-polar mapping method. The algorithm of these methods will first be proposed, and the code will then be generated and be tested on several different omnidirectional images. The images converted will then be compared among each other and be evaluated based on their performance on the resolutions, quality, algorithm used, complexity based on Big-O computations, processing time, and finally their data compression rate available for each of the methods. The most preferable unwarping method will then be concluded, taking into considerations all these factors.

Quaternion Based Thermal Condition Monitoring System

In this paper, we will propose a new and effective machine condition monitoring system using log-polar mapper, quaternion based thermal image correlator and max-product fuzzy neural network classifier. Two classification characteristics namely: peak to sidelobe ratio (PSR) and real to complex ratio of the discrete quaternion correlation output (p-value) are applied in the proposed machine condition monitoring system. Large PSR and p-value observe in a good match among correlation of the input thermal image with a particular reference image, while small PSR and p-value observe in a bad/not match among correlation of the input thermal image with a particular reference image. In simulation, we also discover that log-polar mapping actually help solving rotation and scaling invariant problems in quaternion based thermal image correlation. Beside that, log-polar mapping can have a two fold of data compression capability. Log-polar mapping can help smoother up the output correlation plane too, hence makes a better measurement way for PSR and p-values. Simulation results also show that the proposed system is an efficient machine condition monitoring system with accuracy more than 98%.

Dynamic Threshold Selection for Sequential Learning in Radial Basis Function Networks

For sequential learning in Radial Basis Function (RBF) Networks, there is a requirement for dynamic selection of threshold because a constant threshold is inadequate to accommodate functions of varying amplitudes. In this paper, a new criterion is defined for the dynamic selection of the Euclidean output deviation threshold. Its effect on the learning process experienced by RBF networks with regard to functions of variable amplitude is shown. This improved network can automatically select a suitable threshold for its own supervised learning depending on the objective parameters set to achieve certain accuracy level of the desired output. This paper also proposes further automation to neuron growing and pruning within Radial Basis Function (RBF) neural networks. The proposed dynamic threshold selection technique has shown significant improvement in achieving stable neuron growth rate in dealing with signal amplitude variation.