C. J. Wong

Temporal air quality monitoring using surveillance camera

Various studies showed that inhaled fine particles with diameter less than 10 micrometers (PM10) in the air can cause adverse health effects on human, such as heart disease, asthma, stroke, bronchitis and the like. This is due to their ability to penetrate further into the lung and alveoli. The aim of this study is to develop a state-of-art reliable technique to use surveillance camera for monitoring the temporal patterns of PM10 concentration in the air. Once the air quality reaches the alert thresholds, it will provide warning alarm to alert human to prevent from long exposure to these fine particles. This is important for human to avoid the above mentioned adverse health effects. In this study, an internet protocol (IP) network camera was used as an air quality monitoring sensor. It is a 0.3 mega pixel charge-couple-device (CCD) camera integrates with the associate electronics for digitization and compression of images. This network camera was installed on the rooftop of the school of physics. The camera observed a nearby hill, which was used as a reference target. At the same time, this network camera was connected to network via a cat 5 cable or wireless to the router and modem, which allowed image data transfer over the standard computer networks (Ethernet networks), internet, or even wireless technology. Then images were stored in a server, which could be accessed locally or remotely for computing the air quality information with a newly developed algorithm. The results were compared with the alert thresholds. If the air quality reaches the alert threshold, alarm will be triggered to inform us this situation. The newly developed algorithm was based on the relationship between the atmospheric reflectance and the corresponding measured air quality of PM10 concentration. In situ PM10 air quality values were measured with DustTrak meter and the sun radiation was measured simultaneously with a spectroradiometer. Regression method was use to calibrate this algorithm. Still images captured by this camera were separated into three bands namely red, green and blue (RGB), and then digital numbers (DN) were determined. These DN were used to determine the atmospherics reflectance values of difference bands, and then used these values in the newly developed algorithm to determine PM10 concentration. The results of this study showed that the proposed algorithm produced a high correlation coefficient (R2) of 0.7567 and low root-mean-square error (RMS) of plusmn 5 mu g/m3 between the measured and estimated PM10 concentration. A program was written by using microsoft visual basic 6.0 to download the still images automatically from the camera via the internet and utilize the newly developed algorithm to determine PM10 concentration automatically and continuously. This concluded that surveillance camera can be used for temporal PM10 concentration monitoring. It is more than an air pollution monitoring device; it provides continuous, on-line, real-time monitoring for air pollution at multi location and air pollution warning or alert system. This system also offers low implementation, operation and maintenance cost of ownership because the surveillance cameras become cheaper and cheaper now.

Temporal and spatial air quality monitoring using internet surveillance camera and ALOS satellite image

Air pollution of fine particles with diameters less than 10 micrometers (PM10) is a major concern in many countries due to their ability to penetrate further into our lungs to cause adverse health effects. PM10 also have a significant influence on climate change and visibility. Due to the high cost to set-up air pollution monitoring stations, there are limited numbers of these stations in use. As a result, researchers do not have good temporal development and spatial distribution of the air pollutant readings over a city. This paper is to report upon the usage of an internet surveillance camera to record the temporal development and to map the spatial distribution of air quality concentration. An internet surveillance camera was used to quantify air quality with our own developed algorithm, which is based on the regression analysis of the relationship between measured reflectance components from a surface material and the atmosphere. A newly developed algorithm was applied to compute the temporal development of PM10 values. Advanced Land Observing Satellite (ALOS) images were used to map air quality concentration of the study area. The algorithm was developed based on the aerosol characteristics in the atmosphere to measure PM10 spatial distribution. These PM10 values were compared to other standard values measured by a DustTrak™ meter. The correlation results of both techniques showed that these newly developed algorithms produced a high degree of accuracy as indicated by high correlation coefficient (R2) and low root-mean-square-error (RMS) values.

Total Suspended Solids (TSS) Mapping Using ALOS Imagery over Penang Island, Malaysia

Remote sensing technique is very effective method for water quality mapping through analysis of satellite images over a large coverage of study area. The objective of this study was to test the feasibility of using the ALOS digital image for Total Suspended Solids (TSS) mapping over Penang Island, Malaysia. A new algorithm was developed for detecting and mapping water pollution from the ALOS satellite image. The algorithm used was based on the reflectance model, which is a function of the inherent optical properties of water, and these in turn can be related to the concentration of the pollutants. Water samples were collected using a small boat simultaneously with the acquisition of the satellite image and later analyzed in the laboratory to determine the real concentration of the TSS level. Water sample’s locations were determined by using a handheld GPS. A simple atmospheric correction, namely darkest pixel technique was performed in this study. This is a very simple correction, based on 2 assumption: The first assumption is that in the darkest water pixel of the image there is total light absorption and the radiation light recorded by this pixel comes from the atmospheric path radiance and secondly it is assumed that the atmospheric path radiance is uniform all over the image. The radiation of the darkest water pixel (assumed to represent the atmosphere) is subtracted from the whole image. The darkest pixel is found by searching for the lowest values over water for all wavelengths. The digital numbers for each band corresponding to the sea-truth locations were extracted and then converted into radiance values and reflectance values. The reflectance values were used for calibration of the water quality algorithm. The calibrated TSS algorithm was then used to generate the water quality maps of the study areas. The newly developed algorithm can estimate TSS concentration with linear correlation coefficient square (R) of 0.92. The result obtained indicated that reliable estimates of TSS values for the Penang Island, Malaysia, could be retrieved using this technique.

Development of Migratory Birds Population Monitoring System Using Digital Single Reflex Camera

Traditionally, biologists trace migratory birds by using their naked eye and binocular for counting. The counters will feel tired and easily lost their concentration after waiting and counting the migratory birds for a long period of time. Other than that, when there is a large number of migratory birds fly together, it is very difficult to count them. In this case, the accuracy of the ground count is unreliable. In order to increase the accuracy, we developed a monitoring system for counting migratory birds. Since Digital Single Lens Reflex (DSLR) camera price become more and more affordable, we proposed to use DSLR as a remote sensor to capture images. The images were processed based on the image processing technique to obtain the amount of migratory birds in the image. This software was tested in Kenting National Park during autumn migratory season 2010. The computed populations were compared with the manually counted numbers. The results produce a high degree of accuracy as indicated by the high coefficient of determination (R2) of 0.979. The results show that it can be effective improved the accuracy of ground count.

Aerosol Optical Thickness data retrieval over Penang Island, Malaysia

In this study, we propose a new cost effective approach to retrieve Aerosol Optical Thickness (AOT) data in the visible spectrum by using sky transmittance measurements measured by a handheld spectroradiometer. The transmittance values were measured in spectral region from 350 nm to 1050 nm at the earths surface. The well known Beer-Lambert law was used in this study to retrieve AOT values from the measured transmittance value. The concentrations of particulate matter of less than 2.5 micron (PM2.5) were measured simultaneously with the measurements of the transmittance data. The station locations of the PM2.5 measurements were detemined using a handheld GPS. Three interpolation techniques, namely Kriging Interpolation, Inverse Distance Interpolation and Natural Neighbour Interpolation, were used for mapping the PM2.5 concentration in this study. The accuracies of the three interpolation techniques were determined in this study in order to select the most suitable technique for mapping the air pollution concentration over Penang Island, Malaysia. The results of the analysis indicated that the AOT values were linearly correlated with the PM2.5 readings. AOT and PM10 maps were generated using an interpolation technique (Kriging) based on the measured data. Basically, both PM2.5 and AOT maps agree reasonably well over Penang Island, Malaysia. The highest PM2.5 concentrations were found in densely populated and industrialized areas.

Development of Air Quality Monitoring Remote Sensor Using a Digital SLR Camera

This paper reports that a digital single-lens reflex (SLR) camera can be applied as a remote sensor for monitoring the concentrations of particulate matter less than 10 micron (PM10). An algorithm was developed based on the regression analysis of relationship between the measured reflectance and the reflected components from a surface material and the atmosphere. This algorithm converts multispectral image pixel values acquired from this camera into quantitative values of the concentrations of PM10. These computed PM10 values were compared to other standard values measured by a DustTrakTM meter. The correlation results showed that the newly develop algorithm produced a high degree of accuracy as indicated by high correlation coefficient (R2) of 0.75 and low root-mean- square-error (RMS) of plusmn5 mug/m3.

Improvement on Masking and Flagging Technique on Reducing SST Residual

Sea surface temperature (SST) retrieval depends greatly on algorithm coefficient estimation. However, the algorithm coefficient estimation is relevant to pixel data chosen for multiple regressions analysis. Therefore the proper masking and flagging procedures will ensure that only the good pixels are only chosen for algorithm coefficient analysis. In this paper, new techniques of masking and flagging test were studied. The techniques were based on the characteristics of sea surface temperature at equatorial regions. Statistical techniques included digital satellite data sampling, normality, symmetrical, and homogeneity tests on small-scale and large-scale extraction of group pixels. The SST algorithm was calibrated by correlating in-situ SST data with thermal bands brightness temperatures and sensor zenith angle. In-situ SST data was acquired by drift buoy data, Research on the SEa and islands of Malaysia (ROSES) project and moored buoy data, Malaysia Meteorology Departments. The validation of algorithm with in-situ data showed that combination of of conventional techniques and improved techniques reduced the SST residuals and increase the correlations coefficients.

Calibration of visible and near infrared spectrums for measuring freshness of vegetables

A new nondestructive methods based on optical properties at multiple wavelengths is being applied to measure the freshness of some vegetables. The principle of this method is to determine the absorbance and the reflectance of a sample in visible and near infrared region. When a light beam is illuminated upon a piece of vegetable sample, the majority of the lights penetrate into the sample tissue. Upon entering the tissue, photons scatter in different directions. Some are absorbed, some pass-through to the whole sample and emerge from the opposite side, and some scatter back and reemerge from the region adjacent to the incident center. While the absorption is related to certain chemical constituent of the sample, scattering is influenced by the density, compositions, cells and intercellular structures of samples and therefore can be useful for measuring samples freshness. Our objectives are to investigate the spectral behavior of some vegetables and to develop an algorithm for a non-destructive freshness sensor system using visible and near infrared light sources. The preliminary results of the study showed that the freshness of green mustard leaf and onion using a red (λ = 633 nm) and green (λ = 808 nm) light sources were closely related.

Extracting spatial data from satellite sensor to support air pollution determination using remote sensing technique

Nowadays, air quality is a major concern in many countries whether in the developed or the developing countries. Due to the high cost and limited number of air pollutant stations in each area, they cannot provide a good spatial distribution of the air pollutant readings over a city. Satellite observations can give a high spatial distribution of air pollution. The objective of this study was to map air quality concentration in Penang Island, Malaysia using our proposed developed algorithm from Landsat TM data. The algorithm was developed base on the aerosol characteristics in the atmosphere. PM10 measurements were collected simultaneously with the image acquisition using a DustTrak Aerosol Monitor 8520. The station locations of the PM10 measurements were determined using a handheld GPS. The retrieval of surface reflectance is important to obtain the atmospheric reflectance in remotely sensed data and later used for algorithm calibration. For each visible band, the dark target surface reflectance was estimated from that of the mid-infrared band. The reflectance measured from the satellite [reflectance at the top of atmospheric, p(TOA)] was subtracted by the amount given by the surface reflectance to obtain the atmospheric reflectance. Then the atmospheric reflectance was related to the AOT using the regression algorithm. Similarly, the atmospheric reflectance was related to the measured PM10 values. In this study, the atmospheric reflectance derived from Landsat TM signals were used as independent variables in our calibration regression analyses. The newly developed algorithm produced a high degree of accuracy. The generated PM10 map was also colour coded for visual interpretation and smoothed using an average filter to minimize random noise. This study indicated that the Landsat TM can be a useful tool for air quality study.

Using Image Processing Technique for the Studies on Temporal Development of Air Quality

Nowadays visual information becomes more and more important in almost all areas of our life. This information is represented and processed digitally. Digital image processing is ubiquitous, with applications ranging from television to tomography, from photography to printing, from robotics to remote sensing. In this study, we developed an algorithm to convert multispectral image pixel values acquired by an Internet Video Surveillance camera into quantitative values of concentrations of particulate matter with diameter less than 10 micrometers (PM10). This algorithm was based on the regression analysis of relationship between the measured reflectance components from a surface material and the atmosphere. The newly developed algorithm can be applied to compute the PM10 values. These computed PM10 values were compared to other standard values measured by a DustTrakTM meter. The correlation results showed that this newly develop algorithm produced a high degree of accuracy as indicated by high correlation coefficient (R2) of 7566 and low root-mean-square-error (RMS) values of plusmn3.8306 mug/m3. A program was written by using Microsoft Visual Basic 6.0 to download still images automatically from the camera via the internet and utilized the newly developed algorithm to determine PM10 concentration automatically and continuously. This study indicates that the technique of using Internet Video Surveillance camera images can be a useful tool for monitoring temporal development of air quality.