JPS58212269A - Image pickup element - Google Patents

Abstract

PURPOSE:To attain pickup of data with high resolution while keeping the high speed performance of processing speed, by forming plural photodetecting element groups having different density of arrangement per unit area at an arbitrary position for forming an element plane. CONSTITUTION:In a solid-state image pickup element 1, photodetecting elements E1-En having different photodetecting areas are formed for forming the element plane of the element 1. The total number and the arrangement of the elements are taken arbitrarily, but the photodetecting element groups having at least >=2 kinds of different density areas should be prepared for single element 1. For eample, in a rectangular element plane, the photodetector element groups where the photodetecting sharing area is decreased to 1/4 each and the density of arrangement is 4 times in elements E1-E4 sequentially from the outside to the center is arranged. In this case, at the center including the element E4, the resolution is 8 times as that of the E1 toward both X and Y directions for attaining 64 times picture elements per unit area. Thus, the processing speed is maintained while keeping high resolution performance at the center.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to improvements in solid-state imaging devices used in image processing devices and the like.

AI processing equipment has been used to control industrial robots.
In *, various types of image pickup devices are used, and their device surfaces are formed by arranging light-receiving elements (corresponding to photodiodes) of equal area and shape at a uniform density. Therefore, although it can be said that the resolution within the element plane is constant at any point, on the other hand, higher resolution cannot be required.

C, C0D, (charge coupled dev
In solid-state imaging devices that do not perform electron beam scanning, such as ice) or MOS type, the resolution is determined by the absolute number of light-receiving elements, so in order to capture the imaged object with higher resolution, it is essentially necessary to increase the number of light-receiving elements. However, increasing the number of elements of the same size increases the size of the entire element, and increasing the number of light-receiving elements for a given element size requires more detailed processing. There is no practical limit.

In addition, unnecessarily increasing the number of light-receiving elements will dramatically increase the number of data to be processed, which will require a lot of time for scanning and data processing, making uniform control difficult. ．． This results in .

In any case, if you try to increase the resolution by using an image sensor in which a conventional light-receiving element with the same area and shape is uniformly scaled down over the entire element, it will lead to a decrease in processing speed.
On the other hand, if the absolute number of light-receiving elements is reduced to speed up processing, the resolution will be lowered. Therefore, it is insufficient as a sensor for monitoring and controlling high-speed moving objects such as industrial robots with high resolution.

The present inventor has developed an object that changes rapidly as described above using V! In image processing methods used for image recognition and/or control, in order to ensure high responsiveness, most of the transferred l-frame data is Since the data will be updated with the new next frame data and become meaningless, there is no need to constantly import the data of the entire screen at high resolution, and the processing system will be paying attention to the We focused on the fact that accurate data in the vicinity and with little time difference is important in achieving high efficiency in image processing.

The present invention was proposed in order to eliminate the above-mentioned drawbacks from the viewpoint of Accordingly, it is an object of the present invention to provide an image sensor that can acquire data with a desired high resolution while maintaining high processing speed.

Another object of the present invention is to provide an image sensor that can obtain high resolution as necessary without using a movable optical system such as a zoom lens, which causes a decrease in reliability and responsiveness of the system. There is a particular thing.

The present invention will now be described in detail with reference to the embodiment shown in FIG. (1) is a MOS type, C, C0D, and cape solid-state imaging device that does not involve electron beam scanning. (-), (E
ρ,<W;)% (L) denotes light-receiving elements each responsible for a different light-receiving area, and forming the element surface of the image sensor (1). Although the total number and arrangement positions of the light-receiving elements are arbitrary, a single photosensitive element (1) is provided with a group of light-receiving elements having at least two or more different density regions.

In FIG. 1, (11i, 3 to
An embodiment is shown in which a group of light-receiving elements whose light-receiving area is 1/4 and the arrangement density is 4 times as large are sequentially arranged from the outside toward the center up to (111,). In this case, the central part including (El) has a resolution 8 times VC in both the X and Y directions, and 64 times more pixels per unit area than the local part.

Figure 2 shows an example in which the element surface is radially divided into 16 sections, and each section has light-receiving elements (EL) to (W)k arranged on concentric circles with the light-receiving area gradually decreasing from the outer periphery toward the center. An example is shown.

In this case, the light-receiving area ratio and arrangement density of each element (Taku) to (Ri) are determined by the spacing of each concentric circle, but the light-receiving equivalent area becomes less favorable toward the center, and they are arranged at the same time. The density shall be increased.

＄3 The figure shows a rectangular element - element 'f': on the left side of the crossing line (It), on the right side (l), and in the middle K
Figure 4 shows an arrangement in which the light-receiving elements of (lit) are divided into three groups, and Fig. 4 shows an arrangement in which the light-receiving elements @t) and $n are arranged in two groups with a diagonal line as the boundary. In addition, FIG. 5 shows an example in which light-receiving elements are arranged in two groups, one on the outside (It) and one on the inside (l is shown.

In these three embodiments, the resolution in the area of the light receiving element 0ρ is twice that of the 00 part in both the X and Y directions, and the number of pixels in the unit area is four times as large. In the comparison between section D and section (Et) in Figure 3, 4 in the X and Y directions.
This increases the number of pixels by 16 times.

Furthermore, FIG. 6 shows an embodiment as a linear image pickup device in which the density of the light receiving elements is set higher toward the center.

Here, the functions of these elements (1) will be explained in detail using an example applied to a specific system.

Figure 7 shows a robot arm (AI) with two joints σJ, (JJi-
, (T.J.

A system (2) is shown for alignment to CI's. The small arm (At) in the figure is the movable main arm (
A, ) is supported. These arms (Avs
(The operation control of ρ is performed on the main arm (A
, ) is controlled solely by feedback control based on careful consideration of the information from the image sensor (El) set above.

Each target tip (T, L) has the same shape, but can be identified by the Vwk marks (It) to σD present near the target.

Also, let α be the movable range of the tip of the main arm (At) that supports the small arm (AJ), and β be the movable range of the tip of the small arm (A,) seen from joint σ.Furthermore, the main arm (A,) The minimum operating distance (operating resolution) of the tip is δ, the small arm (A,
) The minimum operating distance of the tip is μ.

Here, it is assumed that α>β and δ>μ.

In other words, this system (2) is positioned as a system that performs wide-area high-precision positioning due to the cooperation between the wide-area movement characteristics of the main arm (A,) and the fine movement characteristics of the small arm (A,).

Such a system, including the method of target identification, is a simpler version of a very common application.

The main conditions required for achieving high efficiency of the CONO system (2) are high positioning efficiency and shortening of positioning time.

First, the main arm (A,) has already completed the rough positioning, and the target is within the movement range of the small arm (At).
! ! Assuming (the state of the solid line), the image sensor (8)
In order to obtain highly accurate positioning, it is necessary to image the extremely limited visual angle (γO) near the target tip with high resolution.At this time, if you try to capture the recognition mark σρ at the same time, It will be necessary to cover the range of γa, but this range can be captured by a conventional image sensor (not shown) K in which light-receiving elements are arranged at a uniform density, and by the device shown in Figure 1. The case of capturing with (1) is compared below.When capturing with element (1) of the present invention,
It is possible to cover the range of viewing angle (γρ with [Ca2 and (g)D). becomes.

On the other hand, in the conventional element, with the same total number of elements, the arrangement density is uniform, so the resolution in the vicinity of an important target is inferior to element (1). Therefore, if the resolution around the target is improved to the same level, the entire range of γ■ will be captured at that resolution, which will not only result in higher resolution than necessary in the periphery of the image where the recognition mark exists, but will also increase the total number of elements and process It becomes a burden on the system.

In other words, element (1) has a visual angle (γD
It is possible to cover the viewing angle (rρ) with the same total number of light-receiving elements as if the total number of light-receiving elements was almost the same as that of the total number of light-receiving elements. possible.

This function demonstrates the effectiveness of the element (1) of the invention in static operation.

Next, arm (A,), (A,) at position (Tt)
, moves to (T, l, stops, and aligns to eye m (T, 3).The difference from the above operation is that the main arm (A,) moves and stops (positioning). However, since it is newly included and the subsequent operations are as described above, only this operation will be considered.

Here, the system (2) is required to be highly efficient. Increasing the movement speed of the main arm (A,) and shortening the positioning time both provide the arm control system with a larger lift time. This is accomplished by proactively communicating general data to the control system.

That is, in the system (2) where feedback control is performed only by data from the image sensor (S),
This requires the image sensor (8) to have a wider field of view. In terms of this predictive function, an image sensor with a viewing angle of (γa, (γD) is more advantageous than an image sensor with only a field of view (γD). If it takes too much time, you will end up wasting your valuable time and the wide field of view will become meaningless.

Based on the above preconditions, the conventional element and the element of the present invention (
1), the field of view in this operation is the γIr
Like snow, it is not something that can be determined based on a fixed number of factors, and it is desirable to have a wider range of factors, so let's assume that γ−(γself>
γ嘗) 4. Make assumptions and compare the two. With conventional elements, the required resolution at the viewing angle rIVc covers the range of γ, so the total number of light-receiving elements increases even more than the required number during static operation, which slows down the processing speed. Don't try to imitate someone and secure enough time.

On the other hand, the element (1) of the present invention maintains the high resolution of 11 parts by widening only the outer peripheral part ((g)Dl(g)D phase part) where the light receiving elements are arranged at a low density. γ,
It is possible to easily widen the VC field of view to a range of 1,000,000,000,000,000,000,000,000,000,000,000,000,000. Similarly, if the field of view of a conventional element is limited to gamma snow for the purpose of shortening the image processing speed, the field of view will be narrowed and this will not lead to an increase in lead time. Therefore, the difference when the pixel element is dynamically operated is greater than when it is statically operated, and this tendency becomes clearer as the operation speed increases.

Therefore, in the element (1) shown in Figures 1 and 2.6,
- Suitable when the target appears in the P visual field from the direction of the target, first appears at the outermost periphery of the visual field, and gradually moves toward the center.

In other words, after capturing a target, it naturally moves toward a high-resolution area, so as the target approaches, the area around the target is automatically zoomed in, and the target data is gradually displayed. It naturally has properties that can be obtained with high resolution. This characteristic works advantageously for both the image processing system and the arm control system. In addition, if you want to operate at a particularly high speed only in a certain movement direction (for example, σD → σρ), the element of the embodiment shown in FIG. 3.4.5 (
1) also exhibits the above properties.

Above, we have explained the system that operates in the simplest plane as an example, but these functions are also effective for 113-dimensional operation as they are.
The usefulness of 1) will increase.

For reference, here we have shown an example in which the camera element side has moved.
The image processing system captures relative changes (movements), so the characteristics described here apply when the image sensor is fixed and the object changes (movement).
Of course, it can be applied in exactly the same way to applications in which the device moves (moves) or moves together.

The scanning method shall be appropriately determined in consideration of the number of light-receiving elements i (number of data) in the operating speed element of the imaging target, the capacity of the processing device, and the like. For example, all of the light-receiving elements ωD-CD in the image sensor (1) shown in FIG. Changes may cause a relative delay in image processing speed, making it difficult to respond appropriately. To solve this problem, as a scanning method in the normal state before the object is captured in the field of view, Enmen 1 (lfl,) scans all elements, and the other elements < nρ ~ (Jli,) Until it is captured, it is better to extract only a few units and perform interlaced scanning, or to make sure that when the object moves within the field of view, a change occurs first at the outermost periphery of the field of view. It is actively used to scan only the group of elements (G) D and ΩD under normal conditions, and after capturing changes in the object within that range, it is scanned with a higher resolution Niremen 1. ), (i,), and by using a method such as scanning only ([nJ, (g) a), high responsiveness can be maintained and captured until the object is captured. It is also possible to further clarify the characteristics that are compatible with ensuring high resolution later.

However, all elements ω, )−(111,)t−J
As mentioned above, even if the device performs 3-dimensional scanning, it can operate at a higher speed than the conventional device.

The scanning method for the element (1) shown in FIG. 2.3.4.5 shall also be carried out in accordance with the above method.

In order to show a specific example of scanning, the scanning method of element (1) shown in FIG. 2 is the spiral scanning shown in FIG. 8, and FIG. , (
It is equipped with a C, C, D0 type transfer mechanism on the circumference, and the transfer lock frequency in the circumferential direction is set to be 16 times the transfer lock frequency in the radial direction. Although the depth scanning method shown in FIG. 8 is shown, it goes without saying that other known scanning methods may be used. Scanning VcX -Y addressing method etc. is not used, but when a self-scanning function is provided as in this example, potential wells (potential wells) due to (3) differences in product, shape, etc. of each light receiving element are used.
It is desirable to standardize the signal before passing it to the transfer mechanism to avoid blooming due to overflow of the ial well.

[Brief explanation of the drawing]

The drawings show embodiments according to the present invention.
The figure is a front view of the surface imaging device, FIG. 6 is a front view of the line imaging device, FIG. 7 is a schematic configuration diagram showing an example of application to a robot system, and FIGS. FIG. 2 is an explanatory diagram showing a scanning method and mechanism. (1)... Solid-state image sensor, (2)...
Robot system, (E,)-ωD... Light receiving element, (BL)...B, B, D, Transfer mechanism, (CL)...C, C, , D, Transfer mechanism, (8)... Image sensor, Patent applicant: Rindo Nakagomi, Patent attorney: Go Hatayama - No. 5
Figure 1 1 Figure 6 H2 E1 Figure 7 γ. Figure 8 Figure 9 Procedural amendment (voluntary) July 1982 zl, B Commissioner of the Japan Patent Office Kazuo Wakasugi 2, title of the invention Imaging device 3, relationship to the case of the person making the amendment! Person's address (residence) 4. Agent (attached sheet) Page 7 of application for detailed statement [Line 7-8 [Light receiving element ("+
)~(E)...'' is replaced by ``Light receiving elenod(-)~
(Ie,)...'', and in the 1st θ line of the same page, ``Each element (Illl) ~ (]1)...'' is replaced with ``Each element (''+) ~ (l[lh)''. In the 12th line of page 13, the phrase ``Reru. Main arm...'' is corrected to ``Reru main arm...''.

Claims (1)

[Claims] 1. In an image sensor having a plurality of light receiving elements,
An image sensor in which an element surface is formed by arranging two or more groups of light-receiving elements with different arrangement densities per unit area or different light-receiving areas at arbitrary positions. 2. The imaging device according to claim 1, wherein the light-receiving element group is arranged so as to be densely arranged from the periphery toward the center. 2 or more light-receiving elements with an arrangement density of 3 iL)
The image pickup device according to claim 1, wherein ¥ is arranged with a line crossing the device surface as a boundary.