JPS5920080A - Picture information processing system using sensor possible for nondestructive readout of signal - Google Patents

Abstract

PURPOSE:To attain high degree picture processing with simple constitution and speed up the processing speed, by reading out repetitively the same picture signal from a photoelectric converting element possible for nondestructive readout of picture signal. CONSTITUTION:A drain, a source and a floating gate are formed on a semiconductor substrate, allowing to obtain image sensors 12, 12' constituted with a photoelectric converting element comprising electrostatic induction type transistors (TRs) S1-SN, S'1-S'N. Coupling lenses 14, 14' are arranged before the sensors 12, 12', and the luminous flux from the lenses 14, 14' is made incident to the sensors 12, 12'. Readout gates G1-GN of the sensors 12, 12' are controlled repetitively with the sequence and combination determined with a common gate control circuit 15, and the rea-out same picture signal is applied to a differential amplifier 16. An output compared differentially at an amplifier 16 is processed with an absolute value circuit 17, a Schmitt trigger circuit 18 and a comparator 23, allowing to attain high picture processing with a simple constitution.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a method for converting the outputs of a plurality of photoelectric conversion elements, which can non-destructively access an image signal a number of times, into a predetermined order through a randomly accessible ``frenzy gate''. The present invention relates to an image information processing method that accesses an image signal and processes image information using its output.

Conventionally, image information processing has typically been performed using, for example, image output from a charge transfer device such as a CCD or BBD, or image output from an image pickup tube. However, since these image outputs are based on the destruction of the photoelectric charge itself, it is necessary to perform a new signal accumulation operation each time the image is read out. For image correlation processing, etc., the image data must be stored in a separate device such as a foot register, and the stored data must be used for processing, which increases the complexity of the device, and when using the accumulated images each time. However, the processing speed was slow, and it was difficult to avoid disturbances such as the movement of objects and changes in lighting.

The purpose of the present invention is to propose a new image information processing method that overcomes the problems in the conventional image information processing methods. From photoelectric conversion elements that can be read out non-destructively,
By repeatedly reading out the same image signal, advanced image information processing is possible with a simple device configuration. Photoelectric devices are the core of such devices.

5 56-150878 etc.
Ansistor) is taken up as an example, and
An example of image information processing is 2 in an automatic focus adjustment device.
Although the image correlation calculation is shown as an example, the present invention is not limited to such foolish embodiments, and can be implemented in any photoelectric conversion element that allows image information to be accessed non-destructively. If this is possible, and if a configuration is used in which the outputs of these elements can be accessed arbitrarily by electrical gates, various other processes in addition to two-image correlation become possible depending on the opening/closing program of the gates.

Hereinafter, the present invention will be described in detail based on the accompanying drawings.

Figure 1 shows a static induction transistor (gate storage type) suitable for use in the image information processing method (Figure 1) of this book.
It is a figure explaining structure 9 action of. 3(a) shows the structure of the same transistor, and FIG. 2(b) shows its equivalent circuit. In FIG. 1(a), in the Si semiconductor substrate constituting the lid static induction transistor, 2 is a drain made of n''Si, 3 is a source similarly made of n+Si,
4 is a floating gate made of p+Si and surrounding sn"Si in a ring. A voltage of, for example, about +7 V is applied to the drain 2, and an appropriate load resistance R is provided to the source 3. The structure is such that a signal is read out from the output terminal O.When the voltage 5 is at a high level, the image is stored and read out, and when it is at a low level, the image is cleared.Such a structure in,
Of the photoelectrons and holes generated in the n- region upon irradiation with light 6, the electrons are absorbed by the drain, and the holes are absorbed by the gate. Therefore, the potential near the gate changes in accordance with the illuminance, and the flowing current changes in accordance with the potential, that is, the brightness of one image. This is shown in the same figure (b
) is shown isometrically. That is, drain 2. Source 3. A static induction transistor consisting of a gate 4 has a gate capacitance C
G, and charges are accumulated there by the incident light 6. Reference numeral 7 denotes a signal reading gate, and the signal is read out via a signal line 8. Due to this structure, the gate of the static induction transistor is controlled according to the amount of incident light, and the drain current becomes a signal to amplify the amount of incident light, and unless the gate potential is cleared,
Each time the reading gate 7 is opened, the signal is read out non-destructively any number of times.

FIG. 2 is a conceptual diagram of an image information processing method using such static induction transistors, and 9 is an image sensor consisting of N static induction transistors St, St...SN, and each Each of the transistors is provided with read gates G+ -Gt, . 11 is 5t-8t...S
This is an image information processing circuit that transmits image information from N through opening and closing of the gate.

FIG. 3 shows an embodiment of the image information processing method according to the present invention. 12 and 12' are N photoelectric conversion elements St-St...SN and SS, such as N electrostatic induction transistors, from which signals can be read out non-destructively. s;...
There are two image sensors including S'N, and an imaging lens 14, 14' is arranged in front of them. Assuming that these imaging lenses and image sensors are arranged apart from each other by a predetermined interval, two images of an object (not shown) will be formed by the imaging lens, and the distance between them will increase as the object distance approaches. As it spreads and the object distance gets farther away, the distance between them gets closer.

If the object is far away, an image is created by the light flux arriving along the solid lines p, , and when the object is close, an image is created by the light flux arriving along the dashed-dotted lines Pt, Pt, and so on. be. Gt-Gt..., GN and Gt, Gt...+GN are the respective static induction transistors Sl-St...SN and S, S2...
This is a gate group for accessing the output of s6. The gate group is arranged in an order determined by the gate control circuit 15;
Opening and closing are controlled according to a predetermined combination.

16 is a differential amplifier that differentially compares the image sensor 12, 12' or the output direction; 17tj: absolute value circuit 5;
This is a Schmidt trigger circuit which outputs a high level output 11111 when the xsFis absolute value circuit output exceeds a predetermined level, and outputs a Fi micro-level signal ao# when it is below a predetermined level. i.e. image sensor-I2 and 12'
If the output of the elements read simultaneously from l"l matches, output #09 is generated, and if they do not match, output of "1" is generated. This is a control circuit that performs control.

Here, the method of two-image phase agony in this example will be explained.

FIG. 4 is a schematic diagram explaining the method.

A and B are image sensors corresponding to image sensors 12 and 12 in FIG.
It is shown as a 7m, 8-element configuration. In order to correlate the two images and finally measure the object distance from the relative positions of the two images, the signals at the symmetrical positions of each image sensor are compared. It is performed as a series of comparison operations including. In other words, in the cycle], the left end of HA and the fifth element from the left end of B are comparatively indicated by 1).Next, A and B are compared with each other, 3 with each other, and 4 with each other. If the two images are located at positions 1 to 4 of A and positions 1 to 4 of B, then
All of these comparison outputs become 0, and the output of the Schmitt trigger 18 becomes 0 all four times. Cycles 2 to 9 below
1, the element outputs of the same numbers shown in A and B are compared. If the same image is formed on A and H, all elements should match and it is cycle 5.
You will know the moment it is executed. In other words, as the distance between the two images increases, the common parts on A and B become smaller.
If there is an identical image portion to the left of the image, it can be seen that the mutual positional relationship of the two images is reversed, and for example, cycle 1 and cycle 9 have this relationship. However, in the example in FIG. 3, when the object is at infinity, the image distance will not be less than the distance between the images due to the light flux represented by the real numbers p+, p;
is a secondary imaging lens, and another imaging lens (D
- When forming a secondary image near the secondary image forming plane on the image sensor, the mutual positions of the two images on the image sensor are reversed as the - secondary image is in front or behind the primary image forming plane. An optical configuration can be realized.

From the above, we know that if the positions and numbers of elements to be compared simultaneously by sensors A and B in each cycle are set as shown in Figure 4, all elements will match when comparing the determined number of elements in each cycle. , the relative positional relationship between the two images (in other words, the 1jll relationship corresponding to the object distance or the focusing state of the imaging lens) can be specified.

Returning again to FIG. 3, the performance of such a process will be explained. The control circuit 2o controls the gate control circuit to open the gates between the elements at the positions as explained in FIG. Sends a gate control command signal to continue. On the other hand and gate 1
9 generates a pulse 27 synchronized with the comparison timing of each element, and furthermore, the It I H signal 2 is always generated.
Give 8 to AND gate 19. If the outputs of the elements to be compared match at the same time as described above, the output of the AND gate 19 becomes "0#" at the timing of the comparison, and if they do not match, the output becomes "IH".

If the number of non-conformities for each cycle is counted by the counter 21, the counter 210 count number is, in principle, 0.
The relative positional relationship between the two images can be determined by specifying the cycle in which In reality, there may be cases of non-O depending on the adjustment state of the optical system, the subject φ, etc., but in any case, the number of non-coincidences in such cases is extremely small. Therefore, the non-conforming count value for each cycle is compared with the output reference value of the output circuit 22, which is a small number close to 0, and when the condition of non-conforming current ≦ marginal value occurs, A cycle counter 2 with a specified cycle number that counts cycle pulses generated one pulse per each cycle, which is separately output from the control circuit 20.
If the output of 4 is latched in the memory circuit 25 and outputted to the data output circuit 26 as appropriate, the output of the data output circuit 26 will correspond to the relative positional relationship of the two images.

As described above, according to the 1ifj image information processing method using the photoelectric conversion element capable of non-destructive readout of the present invention, images can be accessed many times, so that the main data can be accessed by simply opening and closing a gate. Correlation processing between two images as in the example can be performed by analog comparison processing without using a device such as a shift register.

FIG. 3 is a block diagram showing an embodiment of the image information processing method of the present invention, and FIG. 4 is a schematic diagram explaining the two-image correlation processing method. be.

S+-8N: Static induction transistor GI-GN: Read gate 10.15: Gate control circuit 11: Image information processing circuit 16: Differential amplifier 17: Absolute value circuit 18: Schmitt trigger circuit 20: Control circuit 21: Counter 22: Reference value output circuit 23: Comparator 24: Cycle counter 25: Memory circuit 26; Data output circuit

Claims (1)

[Claims] The image signals are accessed multiple times in a predetermined order through electrical gates that can randomly access the outputs of multiple photoelectric conversion elements that allow nondestructive readout of image signals, and the same outputs are used to process image information. An image information processing method characterized by performing the following.