Figure 1 A single module of an LED

Figure 1 A single module of an LED array consisting of 44 lamps with variations in illuminated wavelengths and brightness levels.
The Required Instruments and System Setup
In order to measure the brightness of the LEDs with a high degree of correctness, a USB2000 spectrometer (Ocean Optic Inc., USA) was employed. The spectrometer is fairly flexible to be used as it comes with a fiber optic probe to connect between the light sources and its internal sensor. To be able to move the fiber optic probe to a specific position of the LED array in this application, an XY scanner needs to be used. Alternatively, for an application that requires a moderate degree of correctness and only brightness is required, the captured image from a digital camera can be used for analysis. For the latter system setup, the captured image then serves 2 purposes: for LED position analysis and for measuring the brightness at the center of the LEDs. In our implementation, the latter system setup was selected as it eliminates the requirement to install an additional XY scanner.
The proposed approach relies on acquiring the RGB image of an LED array. During the course of our experiments, all of the still images were acquired using a single CCD (Coupled-charge Device) color camera (XC-711, Sony, Japan) with a zoom lens of 12.5 - 75 mm focal length (C6Z1218, Cosmicar), an image grabbing and processing board (Meter, Matrox Inc., Canada), and a personal computer (Intel Pentium 4 processor, 2.4 GHz). The CCD color camera was fixed at a position of 1 m above an LED array sample and was set to focus on the surface of the sample. A visible full spectrum light source used during our experiments was the GE Sunshine Starcoat® T8
(F15T8/SUN 6PK) light bulb. It was installed to illuminate from a distance of 1 m away with an angle of 45 degree with respect to an LED array. All the ambient light effects were eliminated by enclosing the visual system within a black box. The Matrox Imaging Library (MIL 7.5, Matrox Inc., Canada) was linked to the programs to grab RGB color images of 760 X 760 pixels. The CCD camera was employed for image acquisition with 4,150 lx and F4.0 opening (iris diaphragm). Images were stored in the hard drive from the format of the camera into JPEG file format.
For a single calibration, 2 different images are taken of a single LED array. The first image is taken under 2 conditions: (1) all LEDs are turned off and (2) the external light source is turned on. It is used to automatically locate the LED positions with the approaches to be presented in the next section. The second image IS taken while all LEDs are turned on at their highest brightness while the external light source is turned off. This image is used for brightness measurement.
The Image Processing Routines
The software tool used during the course of our development of the proposed automatic system was Matlab 7.0. Our preliminary studies gave rise to the conclusion that the standard image processing routines [17] successfully applied to similar problems; i.e. object recognition, detection or colour segmentation, failed to locate all the LED positions. This could perhaps result from the variety of colours of the LED packages. In addition, as the normal LED packages are partially transparent (as clearly seen in Figure 1), it results in failures to segment between objects and objects and background. The most difficult task we encountered was that the light gray package LEDs were likely to be removed from the transformed image. As a result, the standard image processing routines fail to locate this group of LEDs.
Figure 2 presents the proposed image processing routines that have been successfully tested on the images of LED arrays. With respect to Figure 2, the input to the proposed image processing routines is the RGB image of an LED array. Two independent processes are then applied to the input image. The processes on the left side are responsible for extracting the light gray package LEDs which are likely to be removed by utilizing the standard image processing routines. The right side processes are responsible for extracting other package colours. Both resulting images are converted to black and white versions and all the internal holes are filled. It is worth noting here that the internal holes of LEDs are the result of the semi-transparent nature of their packages. That is to say we can partially look through the package to see the inside components. These components together with the response of the round shape package to the light source make the inside package colour somehow differ from the colour at the edge.
Both intermediate black and white images are then added together in order to produce the resulting image which seems to grab all the LEDs in the array. The image is then inverted to be ready for a boundary detection process which gives rise to a set of points at the boundaries of all LEDs; B(i) where i is the index of an LED. The final process retrieves a set

ผลลัพธ์ (ไทย) 1:

รูป 1 A โมดูเดียวของอาร์เรย์ LED โคมไฟ 44 โดยการเปลี่ยนความยาวคลื่นแสงและความสว่างระดับประกอบด้วยเครื่องมือที่จำเป็นและตั้งค่าระบบเพื่อวัดความสว่างของไฟ Led มีระดับสูงของความถูกต้อง สเปกโตรมิเตอร์ USB2000 (โอเชี่ยนออปติ อิงค์ สหรัฐอเมริกา) ถูกจ้าง สเปกโตรมิเตอร์มีความยืดหยุ่นค่อนข้างที่จะใช้ ตามที่มันมาพร้อมกับโพรบใยแก้วนำแสงเชื่อมต่อระหว่างแหล่งกำเนิดแสงและเซนเซอร์ภายใน เพื่อให้สามารถย้ายโพรบใยแก้วนำแสงไปยังตำแหน่งเฉพาะของเรย์ LED ในโปรแกรมนี้ ต้องใช้สแกนเนอร์เป็น XY อีกวิธีหนึ่งคือ สำหรับโปรแกรมประยุกต์ที่ต้องการในระดับปานกลางของความถูกต้องและความสว่างเท่า ที่จำเป็น ภาพที่จับจากกล้องดิจิตอลสามารถใช้สำหรับการวิเคราะห์ สำหรับการตั้งค่าระบบหลัง ภาพที่จับแล้วทำหน้าที่ประการที่ 2: สำหรับ LED ตำแหน่งวิเคราะห์ และวัดความสว่างของไฟ Led ในการใช้งานของเรา การเซ็ตอัพระบบหลังถูกเลือกลดข้อกำหนดในการติดตั้งสแกนเนอร์ XY ที่เพิ่มเติมThe proposed approach relies on acquiring the RGB image of an LED array. During the course of our experiments, all of the still images were acquired using a single CCD (Coupled-charge Device) color camera (XC-711, Sony, Japan) with a zoom lens of 12.5 - 75 mm focal length (C6Z1218, Cosmicar), an image grabbing and processing board (Meter, Matrox Inc., Canada), and a personal computer (Intel Pentium 4 processor, 2.4 GHz). The CCD color camera was fixed at a position of 1 m above an LED array sample and was set to focus on the surface of the sample. A visible full spectrum light source used during our experiments was the GE Sunshine Starcoat® T8 (F15T8/SUN 6PK) light bulb. It was installed to illuminate from a distance of 1 m away with an angle of 45 degree with respect to an LED array. All the ambient light effects were eliminated by enclosing the visual system within a black box. The Matrox Imaging Library (MIL 7.5, Matrox Inc., Canada) was linked to the programs to grab RGB color images of 760 X 760 pixels. The CCD camera was employed for image acquisition with 4,150 lx and F4.0 opening (iris diaphragm). Images were stored in the hard drive from the format of the camera into JPEG file format.For a single calibration, 2 different images are taken of a single LED array. The first image is taken under 2 conditions: (1) all LEDs are turned off and (2) the external light source is turned on. It is used to automatically locate the LED positions with the approaches to be presented in the next section. The second image IS taken while all LEDs are turned on at their highest brightness while the external light source is turned off. This image is used for brightness measurement.The Image Processing RoutinesThe software tool used during the course of our development of the proposed automatic system was Matlab 7.0. Our preliminary studies gave rise to the conclusion that the standard image processing routines [17] successfully applied to similar problems; i.e. object recognition, detection or colour segmentation, failed to locate all the LED positions. This could perhaps result from the variety of colours of the LED packages. In addition, as the normal LED packages are partially transparent (as clearly seen in Figure 1), it results in failures to segment between objects and objects and background. The most difficult task we encountered was that the light gray package LEDs were likely to be removed from the transformed image. As a result, the standard image processing routines fail to locate this group of LEDs.Figure 2 presents the proposed image processing routines that have been successfully tested on the images of LED arrays. With respect to Figure 2, the input to the proposed image processing routines is the RGB image of an LED array. Two independent processes are then applied to the input image. The processes on the left side are responsible for extracting the light gray package LEDs which are likely to be removed by utilizing the standard image processing routines. The right side processes are responsible for extracting other package colours. Both resulting images are converted to black and white versions and all the internal holes are filled. It is worth noting here that the internal holes of LEDs are the result of the semi-transparent nature of their packages. That is to say we can partially look through the package to see the inside components. These components together with the response of the round shape package to the light source make the inside package colour somehow differ from the colour at the edge.Both intermediate black and white images are then added together in order to produce the resulting image which seems to grab all the LEDs in the array. The image is then inverted to be ready for a boundary detection process which gives rise to a set of points at the boundaries of all LEDs; B(i) where i is the index of an LED. The final process retrieves a set