TW202438871A - Automatic optical re-inspecting system for multiple re-inspection and multiple re-inspection method - Google Patents

Abstract

The present invention discloses an automatic optical re-inspecting system for multiple re-inspection. The system receives a detection signal and a corresponding object to be tested from the outside. It includes a light source device, an image capture device, and a defect screening module. The light source device provides multiple light sources to the object to be tested in the detection area. The image capture device captures the corresponding defective area on the object to be tested based on the detection signal, obtaining a first image and a second image. The defect screening module determines whether the first image is a true defective image, generates a first replication result, and determines whether the second image is a true defective image based on the first replication result, producing a second replication result.

Description

Automated optical re-examination system for multiple re-examination and multiple re-examination method

The present invention relates to an automated optical review system, and in particular to an automated optical review system for multiple review.

Current optical inspection equipment, including those used in various stages of semiconductor, PCB, LCD and other manufacturing processes, is equipped with re-inspection equipment. Its purpose is to display the defect images obtained through the inspection to personnel for repeated inspection through another independent display device (re-inspection display equipment) for the purpose of re-inspection, so as to reduce the over-inspection and under-inspection rates.

With the increase in labor costs, the traditional one-person-one-machine manual image re-inspection method has too high labor costs. However, if re-inspection is not adopted and only higher-specification optical shooting methods or detection algorithms are used to solve the problems of "first inspection" over-inspection and missed inspection, the inspection time of "first inspection" will be greatly increased, which cannot meet the needs of industrial mass production.

The main purpose of the present invention is to provide an automated optical review system for multiple re-inspections, which receives a detection information and an object to be tested corresponding to the detection information from the outside. The system includes a light source device, an image capture device, and a defect screening module. The light source device provides a plurality of light sources to the object to be tested within a detection range. The image capture device photographs a corresponding defect area on the object to be tested based on the detection information to obtain a first image and a second image. The defect screening module determines whether the first image is a real defect image and generates a first review result, and determines whether the second image is a real defect image based on the first review result to generate a second review result.

Another object of the present invention is to provide a multiple re-inspection method, including: receiving a detection information and a test object corresponding to the detection information from the outside; providing a first light source to the test object in a detection range, and having an image capture device photograph a corresponding defect area on the test object according to the detection information to obtain a first image and generate a first re-judgment result; and providing a second light source to the test object in the detection range, and having an image capture device photograph the corresponding defect area on the test object according to the first re-judgment result to obtain a second image, and judging whether the second image is a real defect image to generate a second re-judgment result.

The present invention adopts a non-online architecture. The separate re-inspection architecture of the present invention only receives the initial inspection information of the third-party AOI equipment from the outside for re-inspection. It is compatible with the inspection information of various specifications of automatic optical inspection equipment and obtains unified re-inspection images and inspection information. In addition, the re-inspection process of the present invention, which combines white light with fluorescence, can improve the accuracy of re-inspection compared to the re-inspection process of the previous technology that only includes one white light or fluorescent light. In addition, the present invention only needs to perform fluorescence inspection on the defects filtered by the white light re-inspection procedure, which greatly reduces the number of fluorescence re-inspections. Compared with the known re-inspection equipment that must perform fluorescence re-inspection one by one according to all defect candidates in the inspection list (the time required for fluorescence inspection is about several times that of white light inspection), the present invention can effectively save the time of the overall inspection. In addition, the control device can control the optical output characteristics of the light source device and the external parameters and/or internal parameters of the image capture device according to the detection information, and strengthen the defect characteristics in the defect area image, so that the defect screening module can perform the re-judgment procedure more accurately.

The detailed description and technical content of the present invention are described below with reference to the accompanying drawings. Furthermore, the drawings in the present invention are not necessarily drawn in accordance with the actual scale for the convenience of explanation, but may be exaggerated. The drawings and their scales are not intended to limit the scope of the present invention.

The present invention proposes an automated optical re-inspection system for multiple re-inspections, which can be used, for example but not limited to, for inner layer inspection of circuit boards, and can be used in conjunction with artificial intelligence (AI) for inspection to implement a re-inspection procedure. The automated optical re-inspection system for multiple re-inspections of the present invention includes, but is not limited to, a non-online architecture that only receives initial inspection information from the outside and is not physically connected to an external automated optical inspection device (AOI) that performs the initial inspection. The above-mentioned optical re-inspection system can access the inspection information stored in the inspection database at the near end or the remote end; the database can simultaneously store multiple automated optical inspection devices of the same or different specifications, and the variations of these embodiments are not within the scope of the present invention. Therefore, when the re-inspection process is performed, the automatic optical re-inspection system of the present invention will take the defect image of the object to be inspected again according to the optical structure of the automatic optical re-inspection system of the present invention based on the externally received inspection information to normalize the defect image for subsequent judgment on whether the defect in the inspection information is a real defect. In this way, the re-inspection image can be normalized under the initial inspection equipment/system structure of different hardware specifications and inspection environments, so as to facilitate the subsequent multiple re-inspection process. Furthermore, the present invention utilizes the re-inspection function of combining white light and fluorescence. According to the defect position provided by the external automatic optical inspection equipment, a white light inspection process is first performed to initially filter out false defects, and then the defect position after filtering is determined according to the fluorescent inspection process to confirm whether it is a real defect, thereby reducing the problem of over-inspection and missed inspection.

The "device", "module", and "device" described in the present invention may be controlled and executed by multiple devices such as background equipment, processors, controllers, or computers of each device and individual equipment, or may be controlled and executed by a single device or equipment as described above. The number of such devices or equipment is not within the scope of the present invention, and the communication method (such as physical or wireless) or communication protocol therebetween is not within the scope of the present invention, so it will not be described in detail in the present invention. The processor may be coupled to the storage unit and execute the corresponding program, and control the operation of the device or equipment according to the program. Such processors and computers may be, for example, a central processing unit (CPU), or other programmable general-purpose or special-purpose microprocessors, digital signal processors (DSP), programmable controllers, application specific integrated circuits (ASIC), programmable logic devices (PLD), or other similar devices or combinations of these devices, which are not limited in the present invention.

In the present invention, each device can communicate directly or indirectly to exchange data. In one embodiment, the execution of each device can be coordinated and allocated by a central control system (such as PLC). However, the present invention does not exclude that the independent controllers or processors of each device can cooperate with each other to complete the work process. The changes of these embodiments are not within the scope of the present invention.

The following is a detailed description of one of the embodiments of the present invention. Please refer to "Figure 1", which is a block diagram of the automated optical review system for multiple re-inspections of the present invention, as shown in the figure: This embodiment discloses an automated optical review system 100 for multiple re-inspections, which reviews the target object Ak by receiving detection information corresponding to the target object Ak provided by the automated optical detection system A.

The automatic optical inspection equipment A is used to detect defects of the object Ak to be tested, and record the defect coordinates or defect types of the object Ak to be tested. The automated optical inspection system A can be an automated optical inspection equipment (AOI) of any existing specification, which is used to perform image analysis on the object Ak to be tested. The image analysis method can, for example, obtain the type or position of the defect through general image processing methods (such as Gaussian, Fourier, binarization, image subtraction, morphological analysis), or use machine learning, deep learning, etc. to identify or locate the type of defect using a neural network to obtain detection information. Generally speaking, the detection information provided by the conventional automatic optical inspection equipment A after detection includes, for example, the type of defect, defect image, defect location, and/or positioning point location. In one embodiment, the object to be tested Ak is a circuit substrate, and the detection information is related to the inner layer defects of the circuit substrate, which is not limited in the present invention.

The multiple re-inspection automated optical re-judgment system 100 performs re-inspection based on the detection information of the object Ak. In one embodiment, the detection information storage device DB can be simultaneously connected to a plurality of automated optical detection devices of the same or different specifications, and the variations of these embodiments are not within the scope of the present invention.

The multiple re-inspection automated optical review system 100 mainly includes a testing platform B for setting up the test object Ak, a light source device 10 corresponding to the testing platform B, an image capture device 20 corresponding to the testing platform B, a control device 30 connected to the image capture device 20, and a defect screening module 40 connected to the control device 30.

The light source device 10 provides a plurality of light sources to the object Ak under test within the detection range. The "detection range" refers to the region of interest on the object Ak corresponding to the shooting direction of the image capture device 20, such as but not limited to the full-frame region of the object Ak under test, a partial region of the object Ak under test, or a defect position of the object Ak under test corresponding to the detection information, etc. Such embodiments are not limited in the present invention.

The image capture device 20 obtains the detection information corresponding to the object Ak from the detection information storage device DB through the control device 30, and shoots the corresponding defect area on the object Ak to obtain the defect area image. In one embodiment, please refer to "Figure 2", which is a block diagram of the image capture device in the present invention, as shown in the figure: the image capture device 20 includes a moving device 21 and a camera 22 arranged on the moving device 21. The moving device 21 receives the defect coordinates through the detection information, moves the camera 22 to the defect position of the object Ak to be tested, and shoots the defect of the object Ak. In one embodiment, the moving device 21 may be, for example, an X-Y linear stage, a robot arm, or other similar devices, which are not limited in the present invention. In one embodiment, the camera 22 may be, for example, an area scan camera or a line scan camera, which are not limited in the present invention. In the present invention, in order to obtain two images presented based on light sources with different characteristics, the image capture device 20 photographs the object to be tested Ak to obtain a first image when the light source device 10 provides a first light source, and photographs the object to be tested Ak to obtain a second image when the light source device 10 provides a second light source. In one embodiment, the first image includes a red light component image, a green light component image, a blue light component image, or a white light component image; the second image includes a fluorescent component image.

In one embodiment, please refer to "Figure 3", which is a block diagram of the optical structure of the present invention, as shown in the figure: the light source device 10 includes a white light source 11 and an excitation light source 12. The white light source 11 is used to provide white light to the object Ak to be tested in the detection range, so that the image capture device 20 obtains a white light component image of the defect area image, or a red light component image, a green light component image, and a blue light component image; the excitation light source 12 is used to provide excitation light to the object Ak to be tested to generate fluorescence, so that the image capture device 20 obtains a fluorescent component image of the defect area image. In the embodiment using the excitation light source 12, an excitation light filter 23 is disposed on the camera head 22 of the image capture device 20, which is used to filter the excitation light and only pass the fluorescence generated by the object Ak to be tested, so as to detect the object Ak. In one embodiment, the image capture device 20 can be, for example, a full-color camera, which is used to obtain a red light component image, a green light component image, a blue light component image and/or a white light component image of the object Ak to be tested.

In one embodiment, when switching between the white light source 11 and the excitation light source 12, the excitation light filter 23 can be switched by the switching device 24. When the excitation light source 12 is started (the white light source 11 is turned off), the excitation light filter 23 is moved to the optical path between the camera 22 of the image capture device 20 and the object to be measured Ak. When the white light source 11 is started (the excitation light source 12 is turned off), the excitation light filter 23 is moved away from the optical path between the camera 22 of the image capture device 20 and the object to be measured Ak. In one embodiment, the object Ak can be transported to the detection range through a conveyor device (not shown) to illuminate and photograph the object Ak. The conveyor device can be, for example, but is not limited to a multi-axis robotic arm, a linear stage, a conveyor belt, a transfer device or other similar devices, which are not limited in the present invention. In addition to the above-mentioned light sources, the light source device 10 may further include RGB light sources, yellow light sources, laser light sources, etc., which are not limited in the present invention; the light source types may include, for example, ring light sources, coaxial light sources, backlight sources, lateral light sources, or adjustable light sources (such as position and angle), etc., which are not limited in the present invention.

The defect screening module 40 is used to determine whether the defect image is a real defect image. In order to perform the inspection efficiently, the defect screening module 40 in the present invention performs multiple re-inspections, first determining whether the first image is a real defect image and generating a first re-inspection result, and then determining whether the second image is a real defect image based on the first re-inspection result and generating a second re-inspection result, thereby reducing the number of secondary inspections. In one embodiment, please refer to "Figure 4", which is a block diagram of the defect screening module in the present invention, as shown in the figure: the defect screening module 40 may include a processor 41 and a storage device 42, and the trained convolution neural network model CV is stored in the storage device 42. The convolution neural network model CV in the storage device 32 is loaded by the processor 31 and the defect image is input to the convolution neural network model CV to determine whether it is a real defect image. The "real defect image" is judged based on the detection information of the automated optical inspection system A. For example, if the detection information of the automated optical inspection system A determines that the corresponding target area has defects, the convolution neural network model CV is trained to analyze whether the defect area image includes defects; if the detection information of the automated optical inspection system A determines that the corresponding target area is a specific defect type (such as dust, outer layer defects, inner layer defects), the convolution neural network model CV is trained to analyze whether the defect area image has defects of the corresponding defect type; thereby determining whether the defect image is a real defect image to complete the image re-judgment. In another embodiment, the convolutional neural network model CV is trained to detect the difference between the target image and the reference image, and to determine whether the target image is a real defect image to generate a re-judgment result; for example, the defect screening module 40 first provides a first light source to the object to be tested Ak, compares the first image with a reference image, and first determines whether the first image is a real defect image to generate the first re-judgment result; then, the defect screening module 40 compares the second image with the reference image based on the first re-judgment result, and determines whether the second image is a real defect image to generate the second re-judgment result.

The "reference image" may be, for example, but is not limited to a master image or a design image, such as a computer-aided manufacturing drawing (CAM), which is not limited in the present invention.

The aforementioned “comparing the second image with the reference image based on the first re-judgment result” means that only if the defects are determined to be real defect images in the first re-judgment result, the image capture device 20 will capture the second image (fluorescent component image) for the corresponding defect positions, and therefore the defect screening module 40 will re-judge based on the second image (fluorescent component image); if it is determined to be a non-real defect image, the defects at the corresponding positions will not be captured, and therefore the defect screening module 40 will only re-judge the defects determined to be defects, thereby reducing the detection time and the amount of calculation. Because the time required for fluorescence inspection is longer than that for white light inspection, the present invention first uses white light re-inspection to determine and filter out false defects, and then conducts fluorescence re-inspection on the filtered defects. By reducing the number of fluorescence re-inspections, the time for fluorescence re-inspection is reduced. For example, if there are 100 defects in the image, 80 false defects are filtered out by white light, and only 20 are left for fluorescence inspection. Fluorescence is excited to take pictures of these 20 defects, and the defects are determined based on the fluorescent image compared with the reference image. Compared with the original need to re-inspect 100 defects by fluorescence, the overall inspection time is much shorter (the time for white light inspection is very short, and the main time-consuming part is the fluorescence inspection, so the time required to inspect 100 sets of defects by white light plus 20 sets of defects by fluorescence is much shorter than the time required to inspect 100 sets of defects by fluorescence.

Please continue to refer to "Figure 5", which is a schematic diagram of the control block of the image capture device and the light source device in the present invention, as shown in the figure: In order to improve the recognition rate of defects and defect characteristics of the re-inspection system, the control device 30 controls the optical output characteristics of the light source device 10 and the external parameters and/or internal parameters of the image capture device 20 according to the detection information, and strengthens the defect characteristics in the defect area image, so that the defect screening module 40 can perform the re-inspection procedure more accurately.

Specifically, since different defect characteristics are better represented by different types of light sources, after the control device 30 obtains the defect type and defect position in the detection information, the control device 30 adjusts the optical output characteristics of the light source Ak provided to the object to be detected in the detection range according to the defect type, and adjusts the internal parameters or external parameters of the image capture device 20 successively or simultaneously; the "optical output characteristics" may include, but are not limited to, light source intensity, irradiation angle, light type (e.g., diffuse light, parallel light, collimated light, etc.), and the like. direct light, etc.), or spectrum, etc.; the "internal parameters" may be, for example, but are not limited to, the aperture size, focal length, focus position, exposure time, ISO value, color temperature value, saturation, image preprocessing, etc. of the image capture device 20, or other such adjustable Internal parameters; the "external parameters" may be, for example, but are not limited to the shooting position of the image capture device 20, the shooting angle, filter switching, the distance from the object under test Ak, the corresponding positional relationship with the light source, etc., or other similar external parameters.

The manifestation of defect types in the optical environment is described below by way of example. However, the following examples are only one possible implementation of the present invention and are not intended to limit the scope of the present invention. This is explained in advance.

If the defect characteristics are easily identified through image processing procedures (such as binarization) in areas with large contrast in hue, saturation, and brightness compared to the surroundings, uniform light (or ambient light) can be provided to the surface of the object under test Ak , so that the brightness of each place on the viewing plane of the object Ak is uniformly distributed. The type of defect characteristics may be, for example, but are not limited to metal discoloration, material surface discoloration, black lines, ink accumulation, leakage of substrate, bright spots, spots, dirt, scratches, etc.

If the defect features have three-dimensional characteristics, lateral parallel light can be provided to the surface of the object to be measured Ak, so that there is an angle between 0 degrees and 90 degrees (but not equal to 0 degrees) between the light path system and the visual plane of the object Ak. and 90 degrees), causing uneven areas in the image to be shadowed. The defective features may be, for example, but are not limited to defects such as vertical lines, knife lines, sanding lines, etc. that cause unevenness on the surface of the object Ak.

The defect characteristics described above are defects inside the object Ak under test or the defects are particularly capable of reflecting light of a specific wavelength. A backlight source can be provided to the back of the object Ak under test, or a light source with a switchable spectrum can be provided. The object under test Ak is illuminated to highlight the defects in the image. The defect features may be, for example, but are not limited to spots (Mura), bright spots, broken bright spots, etc.

In addition to the above disclosed embodiments, the present invention can also combine various light sources to highlight the defect features in the image in accordance with different defect features. The defect area image with the defect features highlighted is finally transmitted to the defect screening module 40 to re-judge whether it is a real defect image.

Please continue to refer to "Figure 6", which is a schematic diagram of the process of the multiple re-inspection method in the present invention (I), as shown in the figure: First, the transport device transmits the test object Ak to be re-inspected to the detection range, and the control device 40 receives the detection information corresponding to the test object from the outside (step S01); then, the control device 40 provides a first light source to the test object Ak on a detection range (step S02); then, the image capture device 20 photographs the corresponding defect area on the test object according to the detection information to obtain a first image (step S03), wherein the first image includes a red light component image, a green light component image, a blue light component image or a white light component image; then, the defect screening module 40 generates a first re-judgment according to the first image. After obtaining the first re-judgment result, the control device 40 provides a second light source to the object Ak within the detection range according to the first re-judgment result, and the image capture device 20 photographs the corresponding defect area on the object Ak according to the first re-judgment result to obtain a second image (step S05), wherein the second image includes a fluorescent component image; finally, the defect screening module 40 determines whether the second image (e.g., the fluorescent component image) is a real defect image, and generates a second re-judgment result (step S06), wherein the area determined not to be a real defect image in step S04 is first filtered out, thereby saving the time for fluorescent detection of non-real defect images, and on the other hand, also saving the calculation amount of the second re-judgment in step S06.

In summary, the present invention adopts a non-online architecture. The separate re-inspection architecture of the present invention only receives the initial inspection information of the third-party AOI equipment from the outside for re-inspection. It is compatible with the inspection information of various specifications of automatic optical inspection equipment and obtains unified re-inspection images and inspection information. In addition, the re-inspection process of the present invention, which combines white light with fluorescence, can improve the accuracy of re-inspection compared to the re-inspection process of the previous technology that only includes one white light or fluorescent light. In addition, the present invention only needs to perform fluorescence inspection on the defects filtered by the white light re-inspection procedure, which greatly reduces the number of fluorescence re-inspections. Compared with the known re-inspection equipment that must perform fluorescence re-inspection one by one according to all defect candidates in the inspection list (the time required for fluorescence inspection is about several times that of white light inspection), the present invention can effectively save the time of the overall inspection. In addition, the control device can control the optical output characteristics of the light source device and the external parameters and/or internal parameters of the image capture device according to the detection information, and strengthen the defect characteristics in the defect area image, so that the defect screening module can perform the re-judgment procedure more accurately.

The present invention has been described in detail above. However, what is described above is only a preferred embodiment of the present invention and should not be used to limit the scope of implementation of the present invention. In other words, all equivalent changes and modifications made according to the scope of the patent application of the present invention should still fall within the scope of the patent of the present invention.

100: Automated optical re-judgment system for multiple re-inspections A: Automated optical inspection system Ak: Object to be inspected DB: Inspection information storage device B: Inspection platform 10: Light source device 11: White light source 12: Excitation light source 20: Image capture device 21: Moving device 22: Camera 23: Excitation light filter 24: Switching device 30: Control device 40: Defect screening module 41: Processor 42: Storage device CV: Convolutional neural network model S01~S06: Steps

FIG1 is a block diagram of the automated optical multiple-review system of the present invention.

FIG. 2 is a block diagram of the image capture device of the present invention.

FIG. 3 is a block diagram of the optical structure of the present invention.

FIG. 4 is a block diagram of the defect screening module of the present invention.

FIG. 5 is a schematic diagram of the control block of the image capture device and the light source device in the present invention.

FIG. 6 is a schematic diagram of the process of the multiple re-testing method of the present invention.

A:Automated optical inspection system

DB: Detection information storage device

100: Automated optical re-examination system for multiple re-examinations

Ak: Object to be tested

B: Testing platform

10: Light source device

20: Image capture device

30: Control device

40: Defect screening module

Claims (19)

An automated optical re-inspection system for multiple re-inspections receives a detection information and a test object corresponding to the detection information from the outside, and the system includes: A light source device, providing multiple light sources to the test object in a detection range; An image capture device, photographing a corresponding defect area on the test object according to the detection information to obtain a first image and a second image; and A defect screening module, judging whether the first image is a real defect image and generating a first re-inspection result, and judging whether the second image is a real defect image according to the first re-inspection result and generating a second re-inspection result. An automated optical review system as described in claim 1, wherein the first image includes a red light component image, a green light component image, a blue light component image, or a white light component image; and wherein the second image includes a fluorescent component image. The automated optical review system as described in claim 2, wherein the light source device includes an excitation light source, providing excitation light to the object to be tested to generate fluorescence; wherein the image capture device, based on the first review result, photographs the corresponding defect area on the object to be tested to obtain the fluorescence component image; wherein the defect screening module determines whether the fluorescence component image is a real defect image to generate the second review result. An automated optical review system as described in claim 1, wherein the detection information is generated from an external automated optical inspection device. An automated optical review system as described in claim 1, wherein the defect screening module includes a convolutional neural network model to determine whether the first image and/or the second image is a real defect image. The automated optical review system as described in claim 1, wherein the defect screening module compares the first image with a reference image to determine whether the first image is a true defect image to generate the first review result. An automated optical review system as described in claim 6, wherein the defect screening module compares the second image with the reference image based on the first review result to determine whether the second image is a real defect image to generate the second review result. An automated optical review system as described in claim 1, wherein the detection information includes defect location information and/or defect category information. The automated optical review system as described in claim 1 further includes a control device for controlling the optical output characteristics of the light source device and the external parameters and/or internal parameters of the image capture device according to the detection information to enhance the defect characteristics in the defect area image. An automated optical review system as described in claim 9, wherein the adjusted optical output characteristics of the light source device include light source intensity, illumination angle, or spectrum. An automated optical review system as described in claim 1, wherein the automated optical review system is a non-online architecture. An automated optical re-interpretation method for multiple re-inspections includes: Receiving a detection information and a test object corresponding to the detection information from an external device; Providing a first light source to the test object in a detection range, and using an image capture device to photograph a corresponding defective area on the test object according to the detection information to obtain a first image and generate a first re-interpretation result; and Providing a second light source to the test object in the detection range, and using an image capture device to photograph the corresponding defective area on the test object according to the first re-interpretation result to obtain a second image, and determining whether the second image is a real defective image to generate a second re-interpretation result. An automated optical review method as described in claim 12, wherein the first image includes a red light component image, a green light component image, a blue light component image, or a white light component image; and wherein the second image includes a fluorescent component image. In the automated optical review method as described in claim 12, if the detection information indicates that the defect of the object to be tested is in an area with a larger contrast in color tone, saturation, and brightness relative to the surroundings, uniform light or ambient light is provided to the surface of the object to be tested. In the automated optical review method of claim 12, if the detection information indicates that the defect of the object to be tested has three-dimensional characteristics, side parallel light is provided to the surface of the object to be tested. In the automated optical review method of claim 12, if the detection information indicates that the defect of the object to be tested is an internal defect, a backlight source is provided to the back side of the object to be tested. In the automated optical review method as described in claim 12, if the detection information indicates that the defect of the object to be tested is particularly capable of reflecting light of a specific wavelength, a light source with a switchable frequency spectrum is provided to illuminate the object to be tested. An automated optical review method as described in claim 12, wherein when photographing the object to be tested, the control device controls the optical output characteristics of the light source device and the external parameters and/or internal parameters of the image capture device according to the detection information to enhance the defect characteristics in the image of the defect area. An automated optical review method as described in claim 18, wherein the adjusted optical output characteristics of the light source device include light source intensity, illumination angle, or spectrum.

Images