JPS6067873A - Object discriminating system - Google Patents

Abstract

PURPOSE:To exclude erroneous discrimination caused by cloud and sea clutters, by detecting the horizon from a searching scene, and extracting a searching object in the vicinity thereof as a real object. CONSTITUTION:When a searching scene is stored in frame memory 10, an image at every frame is scanned along each vertical line by a differentiation circuit 13 and a large differentiated value is outputted to an integration circuit 14 at the horizon part. In the integration circuit 14, a sum is calculated at every horizon and the horizon proposed showing the max. integrated value in each frame is selected in a horizon detecting circuit 15. The estimate horizon is calculated with due regard to a weather condition and a flight in a linear regression operation circuit 16 and only an object present in the vicinity of the estimate horizon in object information, which is preliminarily imparted from an object detecting circuit 11, is set as a real object in an object discriminating circuit 12.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a target identification system used, for example, in an image processing guidance system for anti-ship missiles.

Conventionally, for example, an image processing guidance device for an anti-ship missile has a target identification function that searches for and identifies a target ship. Such a target identification method usually extracts only information that is determined to be a target from each frame-by-frame image of a photographed search scene.

However, with conventional target identification methods, if the detected target image contains g-clutter (unnecessary signals reflected from the sea other than the target), clouds, or a ship different from the target ship, these are removed. The problem was that the ability to identify the true target was poor and the true target could not be reliably identified.

The present invention was made in view of the above-mentioned circumstances, and its purpose is to comprehensively identify the true target from significant information of the entire search scene in an image processing guidance system for anti-ship missiles. An object of the present invention is to provide a target identification method that can reliably identify targets.The present invention uses the horizon line as significant information of the entire search scene, and determines that targets detected outside the vicinity of this horizon are not true targets. Remove. The means for detecting the horizontal line includes a differentiation circuit that differentiates the image of each of the seven frames in the vertical direction, and a horizontal line detection that integrates the differential value of each frame in the horizontal direction and detects the horizontal line based on the integrated value. Consists of means.

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a target identification method according to an embodiment. In Figure #1, 10 is a frame memory, which stores a search scene including a target object (target ship) for each frame. The target detection circuit 11 detects targets (including g-clutter, etc.) for each frame output from the frame memory 10 and outputs the detection results to the target identification circuit 12. On the other hand, the differentiation circuit 13 differentiates each frame image output from the frame memory 10 by scanning a differentiation operator along each vertical line.

In this case, the image is a binary signal quantized into multilevel brightness. The integrating circuit 14 calculates the sum of the differential values within one frame outputted from the differentiating circuit 13 along each horizontal line.

Further, the horizontal line detection circuit 15 selects the one with the maximum absolute value of the integral value obtained by the integrating circuit 14, and if that value exceeds a predetermined threshold value, the horizontal line giving that value is determined as the horizontal line detection value. do.

The linear regression calculation circuit 16 plots the horizontal line detection value in each frame determined by the horizontal line detection circuit 15 and the angle of the center position of the frame field of view in two-dimensional coordinates according to the parallel movement of the field of view due to scanning of an imaging unit (not shown). minimum 2
Performs linear regression calculation using multiplication. The above target identification circuit 1
2 removes targets detected outside the range of a certain area above and below the horizontal line estimated by the linear regression calculation circuit 16, and finds the true target from among the target information provided from the target detection circuit 11. Extract and output.

The operation of the above configuration will be explained. Assume now that a search scene (original image consisting of multiple bits) as shown in FIG. 2(a) is stored in the frame memory IO. In such an image, for example, at the boundary between the sky and the sea, the vertical line brightness changes drastically, as shown in FIG. 2(b). Here, when the image of each frame is differentiated by the differentiating circuit 13 by scanning a differentiating operator along each vertical line, the differentiated value (absolute value) becomes large and large in areas where brightness changes are large. Ru.

Therefore, in the horizontal line portion, a large differential value (FIG. 2(C)) is output from the differentiating circuit 13 to the integrating circuit 14. The differentiating circuit 13 differentiates along each vertical line on the entire screen as shown in FIG. 3(a), and outputs the differential value as shown in FIG. 3(b) to the integrating circuit 14. The integrating circuit 14 calculates the sum of the differential values from the differentiating circuit 13 for each horizontal line. At this time, since the horizontal line exists on a straight line in the horizontal direction, the cumulative value in the integrating circuit 14 becomes a very large value.

On the other hand, even at local boundaries such as clouds, g-clutter, islands, and ships, the differential value at a single vertical line becomes large, but the cumulative value in the horizontal direction is larger than that at the horizontal line. The horizontal line detection circuit 15 then selects the horizontal line with the maximum integral value M in each frame as a horizontal line candidate, as shown in FIG. 3(c), and sets the integral value to a certain threshold. (
IPsl), that horizontal line becomes the horizontal line detection value N.

The linear regression calculation circuit 16 uses the most accurate horizontal line ( estimated horizontal line)
I'm angry.

Ru. With respect to the estimated horizon line determined as shown in FIG. Extracted as a target (target ship).

As described in detail above, according to the present invention, in the target identification method of the image processing guidance device for anti-ship missiles, the horizon is detected from the search scene, and the detection target existing near the horizon is extracted as the true target. . Therefore, even if the targets detected by the target detection circuit include clouds, sea latch, etc., the target can be reliably identified.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a target identification system according to an embodiment of the present invention, FIG. 2(a) is a diagram for explaining the storage contents of the frame memory in FIG. 1, and FIG. 2(b) , (c
) are diagrams for explaining the operation of the differentiating circuit shown in FIG. 1, and FIGS. FIG. 4 is a diagram for explaining the operations of the linear regression calculation circuit and target identification circuit of FIG. 1. 10... Frame memory, 11... Target detection circuit,
12... Target identification circuit, 13... Differentiation circuit, 14.
...Integrator circuit, 15...Horizontal line detection circuit, 16...
Linear regression calculation circuit.

Claims (1)

[Claims] A frame memory that stores the target search scene frame by frame, a differentiation circuit that differentiates the image of each frame outputted from this frame memory in the vertical direction, and a differentiation circuit that differentiates the differential value of each frame obtained by this differentiation circuit in the horizontal direction. horizontal line detection means for detecting a horizontal line based on the integral value;
A target identification method comprising target identification means for identifying a detection target located within a predetermined range with the horizon as a reference based on the horizon detected by the horizon detection means as a true target.