JPS6017382A - Discriminating device of dangerous article - Google Patents

Abstract

PURPOSE:To discriminate whether an object in hand baggage is safe or dangerous in accordance with the size by comparing the number of image regions with a preliminarily set threshold indicating that the object in hand baggage is dangerous. CONSTITUTION:An image region number measuring circuit consists of an image signal correcting circuit 5, an image region count table 6, a counting-up circuit 7, and a calculating circuit 8 and obtains a true number of image regions. First, an image (j) with a region number added which is outputted from a region number adding circuit 2 and a signal j' indicating connection to a region are inputted to an image signal correcting circuit 5 in picture element units. In this circuit, for example, if image signal ''2'' and image signal ''1'' indicate a region ''1'', image signal ''2'' is converted to ''1'' by connection signal ''2''. This corrected image signal is inputted to a count table 6 in picture element units. The number of picture elements in each of image regions (1-l) is counted by the counting-up circuit 7. This picture element number data (h1-hl) is sent to the calculating circuit 8, and a number (k) of image regions where the number of picture elements is not 0 is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a dangerous object determination device that uses image processing technology to determine, for example, whether or not a dangerous object is contained in baggage.

[Technical background of the invention and its problems]

In recent years, after violent attacks by extremists and other groups have begun to occur, airports have begun screening passengers' carry-on baggage with X-ray fluoroscopy equipment and metal detectors to ensure that it does not contain hazardous materials. The X-ray fluoroscope irradiates the baggage with X-rays and obtains a gray scale image according to the radiation absorption coefficient distribution of the baggage.

Conventionally, humans have used this X-ray fluoroscope to visually examine the fluoroscopic image of baggage displayed on a monitor to determine whether it is dangerous or not. Since this judgment is carried out on a large number of pieces of baggage in a short period of time, it is mentally and physically taxing for humans, and there has been a strong desire to automate dangerous goods judgment.

Furthermore, metal detectors detect only the presence of metal in baggage using magnetic lines of force. Therefore, since cameras and pistols in baggage are similarly judged as metal objects, it is difficult to distinguish between dangerous goods and safe goods, resulting in inefficient inspection of dangerous goods in baggage.

[Purpose of the invention]

An object of the present invention is to provide a device that automatically determines dangerous substances in baggage.

[Summary of the invention]

In the case of determining whether a dangerous substance exists in baggage, etc., the present invention uses a perspective gradation image obtained by irradiating the baggage with X-rays or the like to a predetermined density level range [Dl, D2].
The data is converted into a binary signal inside and outside the device. Next, we assigned area numbers to this binary image and at the same time examined the connectivity of this image area.

Then, the true number m of image areas is determined by adding the connectivity of the areas to this image area number. This number of image areas m is compared with a preset "threshold value M indicating the danger of objects in baggage", and depending on the size relationship, it is determined whether the object in baggage is safe or dangerous. It is for making judgments.

〔Effect of the invention〕

According to the present invention, since it is possible to automatically determine the dangerous substances in baggage, the processing time is shorter than in the past, and the burden on humans is reduced.

[Embodiments of the invention]

Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

Generally, when determining the danger of objects in baggage, it is said that ``objects in baggage that can be judged to have a complex internal structure in a fluoroscopic image and a complex density distribution are considered to be of low risk.'' I know the facts. This is because cameras, radios, cassette tape recorders, etc. that are easily carried in hand luggage contain electronic circuit parts, etc.
This is because it can be said that there are almost no dangerous objects with complicated internal structures, and it can be said that dangerous objects such as kitchen knives, knives, scissors, screwdrivers, etc. have simple internal structures. In the present invention, in order to measure the complexity of a grayscale image, a predetermined density level range exists, for example, an interval [DI, D2) in which the average density of a grayscale image of an object has a certain width. The number of density regions is determined as the evaluation value.

FIG. 1 is a configuration diagram of an embodiment of the present invention, and FIG. 2 is a diagram showing a processing procedure of an embodiment of the present invention.

(1) in FIG. 1 is a data conversion circuit. A fluoroscopic image obtained by applying X-rays to baggage in which it is to be determined whether there is a dangerous object inside, for example (512 pixels) x (512 pixels)
The image signal of 8-bit gradation (256 levels l@) is input to the data conversion circuit (11). (In the second section) This data conversion circuit (11) converts the perspective gradation image input in pixel units to the third section. This circuit converts and binarizes data according to the data conversion table shown in the figure.

The data conversion table at this time is composed of readable and writable RAM, and the data conversion table is composed of a readable and writable RAM.
D2] to n-256). (Fig. 2 @) According to this data conversion table, in the input grayscale image, the density area between the density D1 and the density D2 is 0'', and the density area not between them is ``1''''. In this embodiment, the solid part of the object is "0#", and the spatial part of the object and the background are 1. (FIG. 2(C)) FIG. 1(2) is a region numbering circuit, which has a buffer for storing image data of horizontal lines. The area numbering circuit (2) finds an image area by sequentially searching the horizontal line data of the image signal binarized by the data conversion circuit (1). This is done for all horizontal lines, and a different number is assigned to each area that exists as a cluster in the image.

(FIG. 20) FIG. 4 is a diagram showing how the numbers in this area are assigned. For example, for a cluster of "1" in the image signal input to the area numbering circuit (2) in the first column in the internal direction, area numbers R, R+1. ..., R+α has already been assigned, and the area is SRl5RHI...lSR+(!A. At this time, the area numbering circuit (2)
Look at column k+1) and search for 1# block Bx.

+) XA31+, Kiφ and XA” R+ i +1
, =φ (FIG. 4(a)). Let the region number of region X be R+1. However, i-0,...
・, search from the smallest one with α.

H) When XAS, +i=φ and XASR+i+1=φ (FIG. 4 (1))) The area number of area X is the value obtained by adding 1 to the largest area number existing up to that point.

ni XASR+, 4φ and XASR+1+1'AF
When φ (Fig. 4 (C)) The region number of region X is R1i, and the region numbered circuit (
2) is stored.

Area numbered circuit (2
) assigns region numbers to a binary image and stores connectivity.

FIG. 1(3) shows a circuit for measuring the number of image areas. This is the fifth
As shown in the figure, this circuit is composed of an image signal correction circuit (5), an image area count table (6), a count up circuit (7), and a calculation circuit (8), and is used to calculate the true number of image areas. First, the area numbered image j output from the area numbering circuit (2) and a signal jl indicating the connectivity of the area are input to the image signal correction circuit (5) pixel by pixel. In this circuit (5), for example, if the image signal "1" and the image signal '2' are in the area "'1#", the image signal "2" is converted to 1" by the connectivity signal "2'". . This corrected image signal is input to the count table (6) pixel by pixel. Here, the number of pixels in each image area (1 to t) is counted by a count up circuit (7). This pixel number data (h,
~h, ) is sent to a calculation circuit (8), where the number k of image regions whose number of pixels is not 0 is determined. At this time, as shown in Figure 2 (0), the blank area of the image signal excluding the object is area "1", and considering this, the true number of image areas m is m = - 1, but this subtraction is also performed in the calculation circuit (8).(In FIG. 2)) FIG. 1(d) shows the determination result output device. This is the number n of image areas determined by the image area number measuring circuit (3),
This device compares it with a preset "threshold value M indicating the danger of objects in baggage" and outputs the determination result. In other words, if 1) m) M, then the object is safe. 1) If m6M, the object can be judged as a suspected dangerous object. (Figure 2 0) k116
7 are configuration diagrams showing other embodiments of the present invention.

As shown in FIG. 6, this embodiment has a configuration in which an average value measuring circuit (9) is added to the above-described embodiment. FIG. 7 shows a block diagram of this average measuring circuit (9), which is a device for determining the average density of an image area of an object. The data conversion circuit (1) operates in the preset concentration interval (Di,
D2], D2 is set as the average concentration of the object detected in circuit (9), Dl is set to 0, and dangerous substance determination is performed. In FIG. 7, for example, a baggage perspective image signal with 8-bit gradation is input to a comparator α0). This comparator 00) compares this baggage perspective image signal with a preset signal level DO on a pixel by pixel basis, and generates an output signal only when the baggage perspective image signal is larger. The counter 0υ counts the number of pixels whose level exceeds DO among the input perspective image signals. This count value is output as area S.

On the other hand, only those fluoroscopic image signals determined to be thicker than the set value DQ based on the output of the comparator 00) are passed through the AND circuits α. Then, the accumulator (131 inputs only the density values of the fluoroscopic image signal that are larger than the set value DO and accumulates them. This accumulator 03) outputs the density sum of the same area as the area S mentioned above. It has become. In this way, the area S output from the counter α1) and the density sum output from the accumulator (I3) are input to the divider a (a).

This divider α(a) divides the density sum by the area S, calculates and outputs the average density D/S. This average density is a value unique to each object, and the density interval [Dl, D2) set in advance in the data conversion circuit (1) is set to [o].

D/8:], and the same operations as in the first embodiment described above are performed. In this way, in the first embodiment, the interval [
Dx, D2], whereas the number m of image areas changes, one fixed value m is obtained, so stable dangerous object determination can be performed.

Above 1. In the second embodiment, when the background image of the fluoroscopic image signal input to the data conversion circuit is loaded with bias Mi1 variation noise ε, the above-mentioned data conversion table is changed as follows.

In this way, highly accurate dangerous goods judgment can be performed.

In addition, fluoroscopic images of baggage are not limited to X-rays, but can also be irradiated with ultrasonic waves, infrared rays, neutron beams, α, β, >9 magnetic lines of force, etc., and fluoroscopic images obtained by these methods are also applicable to the present invention. Can be used.

In the present invention, a determination is made as to whether or not a single image area is a dangerous object, but the determination can also be made for image areas of a plurality of objects. In this case, binarization is performed using the density level Do, which is lower than the density fluctuation of the perspective image of the object, as a threshold, and multiple image regions that clarify the shape of each object are extracted, and each image region is included. Determine the area you want to use.

If the algorithm shown in FIG. 2 is applied to each region, the determination can be made with higher efficiency.

Furthermore, the present invention can be used not only for fluoroscopic images of baggage, but also for metal surface inspection and the like. For example, 1. Using an X-ray, we can examine both a flat iron plate A and an iron plate B, which is almost the same thickness as this person's but has slight undulations and variations in thickness, and make sure that the thickness of this iron plate is included. concentration interval [D
l, D2), the evaluation value H of the present invention becomes H=0 for iron plate A, and H>O for iron plate B.

2. Iron plate C heated uniformly to temperature T and almost temperature T
However, if we take an image of the iron plate at a temperature of T+2△T only at the defective part using an infrared sensor, and set the density level [D3, D4] corresponding to the temperature range [T- to T, T+△T], we get
The evaluation value H of the present invention is H=O1 for iron plate C and H for iron plate C.
) becomes Q.

In this way, it is also possible to perform judgments for metal surface inspection.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is a diagram showing a processing algorithm of an embodiment of the present invention, Fig. 3 is a block diagram of a data conversion table, and Fig. 4 is a diagram showing the allocation of area numbers. 5 is a block diagram of an image area count table, FIG. 6 is a block diagram of another embodiment of the present invention, and FIG. 7 is a block diagram of an average value measuring circuit. (11...Data conversion circuit, (2)...Area numbering circuit, (3)...Image area number measurement circuit, (4)...
Judgment result output device, (5)...image signal correction circuit, (
6)...Image area count table, (force...count up circuit, (8)...calculation circuit, (9)...
Average value measurement circuit, (1(II... comparator, συ...
Counter, aa...AND circuit, θJ...accumulator, a
4)...Divider. Agent Patent Attorney Kensuke Chika (and 1 other person) No. 1
Figure 3 Figure 4 Ta・f2==eye:2] Figure 5 Figure 6 Figure 7.

Claims (1)

[Claims] A portion consisting of means for obtaining a binary image signal of an object for which it is to be determined whether or not a dangerous substance exists inside, and one block of pixels ``1'' or 0'' of the binary image obtained by this means. A region numbering means for assigning a number to each region and checking the connectivity between partial regions; a partial region number measuring means for measuring the number of partial regions using the results obtained from the region numbering means; A means for comparing the number of partial regions outputted by the region number measuring means with a predetermined threshold value is provided, and an output signal of the comparison means is output as a signal indicating the presence of a dangerous substance in the object. A dangerous goods determination device characterized by: