JPS6312984A - Target detecting and tracking device - Google Patents

Abstract

PURPOSE:To satisfactorily detect and track a target by using function data which becomes maximum, only as for a feature quantity of a known target, using uniform function data as for an unknown feature quantity, and calculating an evaluation value. CONSTITUTION:A video signal outputted from an image pickup device 11 is brought to an area division by an area dividing and labeling circuit 13 through an A/D converting and pre-processing circuit 12, each area which has been labeled is supplied to a feature extracting circuit 14, and a feature is calculated and extracted. A feature quantity which has been obtained by the circuit 14 is supplied to an evaluating circuit 15 and an evaluation function generating circuit 19, as well. The circuit 19 generates an evaluation function data which becomes maximum, and a uniform evaluation function data, as for a known feature quantity, and as for an unknown feature quantity, respectively, and also, corrects successively the uniform evaluation function data to the prescribed evaluation function data, whenever a feature quantity of a target is detected thereafter, and stores it in an evaluation function storing circuit 16. In this way, many features are used, and a range of the feature quantity can be set narrowly, therefore, the target can be detected and tracked exactly.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Industrial Application Field) This invention provides information on the maximum value, minimum value, size, and coordinates of each region in images supplied from a television camera or an infrared imaging device. , summation, average, aspect ratio, perimeter etc. 1! ! II
The present invention relates to a target detection and tracking device that can track a target within its field of view by extracting the target's weight, detecting the target from the weight it carries, and controlling the ms force direction.

(Prior Art) Conventionally, this type of target detection and tracking device is configured as shown in FIG. That is, in FIG. 3, the me device 1
The video signal output from 1 is analog/digital (A
/D) The conversion/preprocessing circuit 12 converts the analog signal into a digital signal, performs smoothing (noise removal), etc., and extracts it as a digital image signal.

This signal is divided into regions by the region division/labeling circuit 13.

Label each area, e.g. 1, 2, 3. A number will be assigned. Each region thus labeled is supplied to the feature extraction circuit 14, and features such as size and coordinates are extracted. The feature amount extracted here is supplied to the evaluation circuit 15.

The evaluation circuit 15 is supplied with an evaluation function that maximizes the target feature amount stored in advance in the evaluation function storage circuit 16, and the evaluation circuit 15 calculates an evaluation value from the evaluation function and the extracted feature amount. be done. From the evaluation values obtained here, a maximum evaluation value detection circuit 17 detects the one with the highest evaluation value from each region, and further extracts an error signal. This error signal is supplied to the servo circuit 18 and the Fill device 1
1 direction is controlled, and thereby the detected target is tracked.

However, in the conventional target detection and tracking system described above, it is necessary to use many features or narrow the range of special weights in order to perform target detection and tracking well. There is a limit to the features, and the goal is to wait! I! It is also difficult to know the amount accurately. In addition, if each evaluation function is initially set to maximize each feature value including the uncertain feature ms, the target may not be detected due to a shift in one feature value, and target detection and tracking may cause problems.

(Problems to be Solved by the Invention) This invention solves the problem that it is difficult for conventional devices to accurately detect a large number of target features, and
Setting each evaluation function that maximizes each feature including fifi improves the problem where the target was not detected due to the deviation of one feature, and when detecting, the known feature of the target can be used within a narrow range. It is an object of the present invention to provide a target detection and tracking device that can detect and track a target well by using the extracted characteristic amount of the target during the detection and tracking.

[Configuration of the Invention (Means for Solving the Problems) That is, the target detection and tracking device according to the present invention has an 11ilF! & section, an extraction section that is supplied with signals from the imaging section and extracts a plurality of feature amounts for each region of the image, a storage section that stores predetermined function data in advance, and storage data from this storage section. and an arithmetic unit that derives an evaluation value for the feature amount from the output data from the extraction unit; and a detection unit that specifies one v4 region of the ms from the output evaluation value of the 8I calculation unit and outputs an error signal. , a control section to which the output signal of the detection section is supplied and controls the direction of the III image of the II image section, and a previously known special wI! Generates function data that maximizes this value for , generates uniform function data for unknown feature quantities, and sequentially generates uniform function data every time the target feature quantity is detected. and a generating section that corrects the function data into predetermined function data and stores the modified function data in the storage section.

(Operation) In other words, the target detection and tracking device with the above configuration uses the function data that maximizes only the characteristic v11 quantity of the known target, and uses uniform function data for the unknown characteristic quantity. Calculate the evaluation value. Thereafter, each time the target feature m1ii is extracted, the linear function data is corrected so that each feature section is maximized, and an evaluation value is calculated for each feature.

(Embodiment) An embodiment of the present invention will be described below with reference to FIGS. 1 and 2. However, in FIG. 1, the same parts as in FIG. 3 are denoted by the same reference numerals, and only the different parts will be described here.

FIG. 1 shows a configuration in which the present invention is applied to the device shown in FIG. Also supplied. This evaluation function generation circuit 19 uses a previously known characteristic! For lI-, evaluation function data is generated such that this is the maximum value, and for unknown feature amounts, uniform evaluation function data is generated, and from then on, each time a target feature amount is detected, evaluation function data is generated. Uniform evaluation function data is corrected to predetermined evaluation function data, and each evaluation function data is stored in the evaluation function storage circuit 1G.

The operation of the above configuration will be explained below.

First, the evaluation function generation circuit 19 generates evaluation function data that maximizes one of the known feature quantities of the target, for example, as shown in FIG. For the unknown feature quantity, uniform evaluation function data is generated, for example, as shown in FIG. 2 <b), and each is stored in the evaluation function storage circuit 16. Here, when the target is detected, the evaluation function generation circuit 19 corrects the evaluation function data as shown in FIG.
Change the evaluation 1iIIIirjA number data stored in 6 to @.

Therefore, detection from the next screen is performed using the corrected evaluation function data.

To explain specifically, first, evaluation function data is generated to maximize one of the target known feature values, for example, contrast, and other unknown features are generated.
The quantity evaluation function data is generated uniformly. Here, when the contrast characteristic W1 amount reaches the maximum, from the next screen, the characteristic l
The evaluation function data is also modified so that the value is always the maximum.

Next, the evaluation function data is corrected so as to maximize one of the unknown feature quantities, for example, the size, and a special extraction of the target size is performed. Furthermore, each time a feature is extracted, other uniform evaluation function data is modified,
Other feature quantities are extracted sequentially. With this, all special features! The maximum value can be obtained for the most.

Therefore, the target detection and tracking device configured as described above uses many features and can set a narrow range of feature amounts, so that it can accurately detect and track a target.

[Effects of the Invention] As detailed above, according to the present invention, when detecting the target's known waiting 'eIL! It is possible to provide a target detection and tracking device that can detect ik in a narrow range and can perform target detection and tracking well by using the extracted target characteristic Wim during tracking.

[Brief explanation of drawings]

FIG. 1 is a block circuit configuration diagram showing an embodiment of the target detection and tracking device according to the present invention, FIG. 2 is a diagram for explaining the evaluation function data generation means of the same example, and FIG. 3 is a diagram showing the conventional FIG. 2 is a block circuit diagram showing the configuration of a target detection and tracking device. 11...l1li image device, 12...A/D conversion/preprocessing circuit, 13...area division/labeling circuit, 14
...Standby extraction circuit, 15...Evaluation circuit, 16...
Evaluation function storage circuit, 17... Maximum evaluation value detection circuit, 1
8... Servo circuit, 19... Evaluation function generation circuit.

Claims (1)

[Claims] an imaging section that outputs an image signal; an extraction section that is supplied with the signal from the imaging section and extracts a plurality of feature quantities for each region of the image; a storage section that stores predetermined function data in advance; a calculation unit that derives an evaluation value for the feature amount from the stored data from the extraction unit and the output data from the extraction unit; and a calculation unit that identifies one region of the image from the output evaluation value of the calculation unit and outputs an error signal. a control unit to which the output signal of the detection unit is supplied and which controls the imaging direction of the imaging unit; and a control unit that generates function data such that the maximum value is obtained for the previously known feature quantity and detects the unknown feature. a generating unit that generates uniform function data for the quantity and sequentially corrects the uniform function data into predetermined function data and stores it in the storage unit every time a target feature quantity is detected thereafter; A target detection and tracking device comprising: