JPS6149285A - Picture processing device - Google Patents

Abstract

PURPOSE:To enable to execute highly accurate picture processing which improves S/N basically by changing a light source position successively and executing plural numbers of pick-up each time an object is picked up. CONSTITUTION:Plural light sources 31-3N are respectively lighted up by a control circuit successively and illuminates an object 1. The light source 32 seems to be installed in an ITV camera, but, this only indicates that the setting position of a light source and a camera is overlapped. Each time light sources 31, 32...3N are lighted up, a video signal obtained through the ITV camera 2 is amplified by a lamp 4, and after that, the signal is made binary by being compared with a prescribed binary level by a comparator 5. Next, the signal is succcessively stored through a switch 6 to picture memories 71, 72...7N. An operation circuit 8 executes a prescribed operation based upon the data stored at memories 71...7N.

Description

[Detailed Description of the Invention], C Technical Field to Which the Invention Pertains] The present invention relates to an image processing device using an imaging device such as an ITV (industrial television) camera, and in particular to an image processing device capable of improving the S/N ratio. The present invention relates to an image processing device.

[Prior art and its problems]

In general, when the contrast between the background and the object is low, or when an ITV camera images an object placed in a background that is likely to cause halation, such as a metal background, the background signal and object signal are separated. is difficult. Conventionally, this has been dealt with as follows.

b) Take an image of the object once and differentiate the image signal (
Laplacian; includes second-order derivatives. ) and other signal processing.

(1) The light source position is fixed, and signal processing is performed such as capturing multiple images and averaging them.

c) Capturing images multiple times by changing the imaging direction.

However, in the above method (a) 1), when the noise becomes large enough to form, it is difficult to distinguish it from the object even if signal processing such as differentiation is performed, and the S/N ratio cannot be improved. Furthermore, if the light source is the same, halation tends to occur at the same position, making it difficult to separate noise. On the other hand, method C) above is based on the same principle that humans would use to change their viewpoint to make it easier to see when an object is difficult to see due to halation. This requires complex processing such as automating focus adjustment, providing a mechanism to move the camera, and adjusting the positions of the images of the objects on the screen because they differ from each other. There is a drawback.

[Purpose of the invention]

This invention was made under these circumstances, and does not require any special mechanism to move the camera, does not require complicated calculations such as Laplacian, and can be easily performed using a simple device.
An object of the present invention is to provide an image processing device that can improve the /N ratio.

[Key points of the invention]

This invention eliminates shadows such as halation by sequentially changing the light source position each time an object is imaged and taking images multiple times, and by moving the camera and taking images. This is to achieve the same effect as eliminating the . In other words, when humans see something dazzling, they can change their perspective and see it by moving the light source in the opposite direction.

[Embodiments of the invention]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a waveform diagram for explaining its operation.

In FIG. 1, 1 is an object to be inspected, 2 is an imaging device such as an ITV camera, 31 to 3N are light sources, 4 is an amplifier, 5 is a comparator, 6 is a switch, 71 to 7N are image memories,
8 is an arithmetic circuit (majority circuit).

Each of the plurality of light sources 31 to 3N is sequentially turned on by a control circuit (not shown) to illuminate the object 1. Note that although the light source 32 appears to be provided inside the ITV camera 2 in FIG. 1, this only shows the case where the set positions of the light source and the camera coincidentally overlap. light source 3
The image (video) signal obtained through the ITV camera 2 every time 1.32...3N is turned on is amplified by the amplifier 4, and then compared with a predetermined binarization level by the comparator 5. The images are binarized and sequentially stored in image memos 1771, 72, . . . , 7N via switch 6. Arithmetic circuit 8 performs predetermined calculations based on data stored in memories 71-7N.

Now, the image signals obtained when the light sources 31 and 32.3N are turned on are shown in FIGS. 2(a) and 2(b), respectively.

As shown in (c), and the binarized images are as shown in (a)′, (b)′, and (c)′ in the same figure, even if the object is the same, the illumination position Due to the difference in (angle), the noise components N will also appear at different positions. In FIG. 2, P is an image signal component representing the object, P' is its binary signal, N' is a binary signal of the noise component N, and VH is a vertical synchronizing signal. Therefore, in the arithmetic circuit 8, for example, the memory 71
Read data from ~7N in parallel and add them until the added value exceeds a predetermined level.

If we assume that the logic is "1" when the noise is generated, the position of noise occurrence will be random and the signal component will repeatedly occur at approximately the same position, so a signal component as shown by S in the same figure (d) can be obtained. , noise components are eliminated and become "O'."

In other words, the signal component is generated repeatedly even if the position of the light source is changed, whereas the noise component is generated only discretely, so the sum of these components is calculated to determine whether it exceeds a predetermined threshold level. By determining whether this is the case, noise can be removed and the S/N ratio can be fundamentally improved. This makes it possible to easily and accurately inspect scratches on metal foil, read printed characters, and inspect patterns on printed boards, which were previously considered difficult, and expand the range of applications. It will also bring some advantages.

〔Effect of the invention〕

According to this invention, since the illumination position is changed each time an image is taken, a mechanism for moving the imaging device is not required, and therefore, a high-precision image with a fundamentally improved S/N ratio can be obtained with a simple configuration. This brings about the advantage that processing becomes possible.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, and FIG. 2 is a waveform diagram for explaining its operation. Description of symbols 1...Inspection object, 2...Imaging device (ITV camera), 31-3N...Light source, 4...Amplifier, 5...
Comparator, 6... Switch, 71-7N... Image memory, 8... Arithmetic circuit (majority circuit).

Claims (1)

[Claims] a plurality of illumination means that sequentially illuminate the object from different angles; and an imaging means that captures images of the sequentially illuminated object;
A binarization means for binarizing the image pickup signal, a plurality of memory means for selectively storing binarized images obtained by sequentially binarizing each of the illumination means, and the contents of each memory means are stored. 1. An image processing device comprising: arithmetic means for sequentially and parallelly reading data and calculating a sum thereof, and removing discretely generated noise components by calculating the sum.