JPH0534762A - Camera with automatic iris function - Google Patents

Abstract

PURPOSE:To provide a camera with an automatic iris function which can scrupulously, smoothly and automatically control a diaphragm. CONSTITUTION:This camera is provided with a lens 101, the diaphragm 102, a diaphragm driving means 103 which adjusts the diaphragm 102, a detection means 104 which detects an image through the lens 101 and the diaphragm 102, a division means 105 which divides the image obtained by the detector into plural areas, an arithmetic means 106 which calculates the brightness of the divided respective areas and a control means 107 which controls the driving means 103 according to the brightness of the respective areas. Besides, the control means 107 is provided with a neural network 108 which relates the brightness of the respective areas with the adjusting degree of the diaphragm.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a camera having an auto iris function, that is, an automatic diaphragm function.

[0002]

2. Description of the Related Art A conventional camera with an auto iris function adjusts an aperture by comparing a brightness level in the whole screen or a part of the screen with a reference level, and therefore, a correction switch is required to cope with backlight or the like. In the new one, the screen is divided into several areas, the brightness level in each area is extracted as the evaluation level, the average, maximum and minimum parameters of the evaluation level are calculated, and many rules are used. The aperture has been adjusted by expressing the priority and weight of each area on the screen of various patterns.

[0003]

However, in such a conventional camera, the rule for determining the priority order becomes complicated as the number of patterns increases, so that the types of screen patterns are limited. However, since it is difficult to smoothly perform interference / complementation between rules, it is difficult to obtain a stable screen especially for a moving subject.

The present invention has been made in consideration of the above circumstances, and provides a camera having a more detailed and smooth camera aperture adjusting function.

[0005]

According to the present invention, a lens and a diaphragm, diaphragm driving means for adjusting the diaphragm, detecting means for detecting an image through the lens and diaphragm, and a plurality of images obtained by the detector are provided. Dividing means for dividing into areas
A calculation means for calculating the brightness of each of the divided areas and a control means for controlling the diaphragm driving means corresponding to the brightness of each area are provided, and the control means sets the brightness of each area and the adjustment degree of the diaphragm. It is intended to provide a camera with an auto iris function, which is provided with a neural network for associating.

FIG. 1 is a block diagram showing the basic configuration of the present invention, in which 101 is a lens, 102 is a diaphragm, and 103 is a diaphragm 102.
Aperture drive means for adjusting the
Detecting means for detecting an image via 102, 105 is detecting means 1
A dividing unit that divides the image obtained by 04 into a plurality of regions, 106 is a calculating unit that calculates the luminance of each divided region, and 107 is a control unit that controls the diaphragm driving unit 103 according to the luminance of each region. The control means 107 is provided with a neural network 108 which correlates the brightness of each area with the adjustment and degree of the diaphragm 102.

[0007]

In FIG. 1, the detecting means 104 detects an image through the lens 101 and the diaphragm 102, and the dividing means 105 divides the image obtained by the detecting means 104 into a plurality of areas. The calculation means 106 calculates the brightness of each divided area,
A neural network 108 associates the brightness of each area with the degree of aperture adjustment. Then, the control means 107 controls the diaphragm driving means 103 according to the degree of adjustment of the diaphragm.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below based on the embodiments shown in the drawings. This does not limit the invention.

In this embodiment, a large number of screens and aperture data considered to be optimal at that time are prepared in advance as learning data. Next, each screen is divided into several areas and the brightness level of each area is extracted.
Calculate some parameters such as minimum and average. The brightness level and each parameter thus obtained are normalized to a numerical value of 0 to 1 and used as input of neural network learning data. As the output data of the learning data, the aperture of each screen is normalized and represented by 0 to 1 for use.

Using the learning input / output data thus created, the neural network is made to learn, and the relationship between the brightness level of each area and the aperture is learned. By inputting information from a new screen different from the learning data to the network after learning, as in the case of learning,
As the output, the normalized value of the aperture required for the screen will be obtained. This network is used for the camera as an auto iris system.

As an embodiment of the present invention, an auto iris system in an 8 mm camera will be described. FIG. 2 shows a block diagram of the system. In FIG. 2, 1 is a lens, 2a is a diaphragm, 2 is a diaphragm driving device, 3 is a CCD for detecting an image through the lens 1 and the diaphragm 2a, and 4 is an AGC circuit for controlling the gain of a video signal output from the CCD ( Automati
c Gain Control Circuit) 5 is a video signal processing circuit for processing the video signal from the AGC circuit 4 and outputting it to a signal recording circuit (not shown). 6 and 7 are bandpass filters,
8 is a low-pass filter, 9 is a time division switching circuit, and 10 is AD
Converter, 11 is a screen division circuit, 12 is a digital integration circuit, 13
Is a microcomputer, 14 to 16 are DA converters, and 17 is a motor drive unit for focusing the lens 1.

The signal photoelectrically converted by the CCD 3 is sent as a video signal to the video signal processing circuit through the AGC circuit 4. Further, the output signal of the CCD 3 is sent to two band pass filters 6 and 7 and a low pass filter 8 to extract a high frequency component for auto focus and a brightness level for auto iris, respectively. The signal that has passed these filters is AD by the AD converter 10.
(Analog to digital) Screen conversion circuit 11 after conversion
Are divided into areas 1 to 3 shown in FIG. 3, and integrated (digitally integrated) by the digital integration circuit 12 for each area.

These filters are switched by the time division switching circuit 9 controlled by the microcomputer 13, and the integration result is an evaluation value indicating the degree of focusing when the bandpass filter 6 or 7 is selected, and the lowpass filter 8 is selected. When is selected, the evaluation value indicates the brightness level of each area.

In this embodiment, six evaluation values (luminance levels) corresponding to the respective areas 1 to 4 when the low-pass filter 8 is selected are input to the microcomputer, the state of the screen is judged, the diaphragm 2a and the AGC circuit 4 are judged. The backlight and over-forward light are corrected by adjusting the reference voltage (reference brightness level) and so on.

Next, FIG. 4 shows a flowchart of the processing performed by the microcomputer 13. Microcomputer
13 receives the evaluation value of each area, its maximum value, minimum value,
Calculate parameters such as the average value. Next, these parameters and evaluation values are each normalized to 0 to 1 and given as inputs to the neural network, and processed to obtain correction values such as the aperture 2a and the reference voltage of the AGC circuit 4 as outputs ( 5), and the form of the network of FIG. 5 may be divided into several networks depending on the time. The neural network is stored in the microcomputer 13 as a program.

According to this embodiment, it is not necessary to construct a complicated rule, and the diaphragm is determined from the brightness level pattern of the entire screen, so that it is possible to cope with the state of the screen such as retrograde / excessive forward light. Also, when a new pattern screen is required, it is much easier to modify than a system created by building rules. The neural network is good at ambiguity processing and complementary processing for inputs other than learning data, so it can smoothly respond to screen patterns that could not be considered at the time of system construction, and there is no gap between rules and rules so movement of A stable screen can be supplied to a certain subject, and smooth screen changes can be expected. In this way, an excellent auto iris function can be enabled.

[0017]

According to the present invention, it is possible to provide a camera with an auto iris function that can automatically control the aperture finely and smoothly.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic configuration of the present invention.

FIG. 2 is a block diagram showing an embodiment of the present invention.

FIG. 3 is an explanatory diagram showing division of a screen.

FIG. 4 is a flowchart showing the operation of this embodiment.

FIG. 5 is an explanatory diagram showing inputs and outputs of the neural network of this embodiment.

[Explanation of symbols]

 1 Lens 2a Aperture 2 Aperture Driving Device 3 CCD 11 Screen Division Circuit 13 Microcomputer

Claims (1)

Claim: What is claimed is: 1. A lens and an aperture, an aperture drive unit for adjusting the aperture, a detection unit for detecting an image through the lens and the aperture, and an image obtained by the detector in a plurality of regions. And a control means for controlling the aperture driving means in accordance with the brightness of each area, and the control means controls the brightness of each area. A camera with an auto-iris function, which is equipped with a neural network that associates the degree of aperture adjustment with the degree of aperture adjustment.