JP2853219B2 - Vehicle detection device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a vehicle detection device that automatically detects whether a vehicle is stopped in a monitoring area by applying image processing technology. It is.

[Prior Art] Conventionally, the following method has been used to determine the stop of a vehicle using a vehicle detection device to which this type of image processing is applied. That is, for each storage frame of the image, a spatial differentiation process is performed to extract the edge of the vehicle, and the degree of overlap (D) of the edge between two consecutive frames is calculated by the following equation. D = (the number of pixels of the overlap portion between the frames of the edge) / (the total number of pixels of the edge of the previous frame) When the overlap D becomes a certain value or more, it is determined that the vehicle is stopped.

[Problems to be solved by the invention]

However, according to this method, it is possible to reliably determine if the edges of the stopped vehicle exist in two consecutive frames so as to substantially overlap, for example, there is a person who gets on and off in front of the vehicle, When a part of the edge of the vehicle disappears, the value of the overlap degree D becomes small, and the stop of the vehicle may not be determined.

The present invention has been made in view of the above points,
It is an object of the present invention to provide a vehicle detection device that can reliably detect a stop of a vehicle even when a part of an edge of the vehicle disappears due to an obstacle or the like in front of the vehicle. Is to do.

[Means for solving the problem]

The vehicle detection device according to the present invention includes: an imaging unit that captures an image of a monitoring area to obtain an image signal; an A / D unit that digitizes an image signal obtained by the imaging unit; Spatial differential means for performing a spatial primary differentiation process to output a differential absolute value, comparing means for extracting an edge image by comparing the differential absolute value with a preset value, and a histogram of the edge image Histogram measurement means for obtaining a main edge by performing a threshold process for each of the horizontal direction and the vertical direction, a previous image memory for storing the main edge of the previous image, and using the main edge Detection mask generating means for generating a detection mask; calculating an exclusive OR of a main edge of a current image and a main edge of a previous image for each of the detection masks; Characterized by comprising a determination means, further, said detecting mask making means,
The main edge is labeled, and a mask obtained by expanding a predetermined pixel for each labeling is used as a detection mask. Neighboring the lengths in both directions obtained by projecting in the horizontal direction
It is characterized in that a square as a piece is used as a detection mask.

[Action]

The vehicle detection device of the present invention obtains a histogram of an edge image, and performs horizontal and vertical threshold processing for each of the main directions.
A detection mask is created for each of the main edges, the previous image and the current image are compared for each detection mask, and when a predetermined number of the plurality of detection masks match, it is determined that the vehicle has stopped. It is.

〔Example〕

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of one embodiment of the present invention. Reference numeral 1 denotes an imaging unit, which is, for example, a television camera using a CCD imaging device or a vidicon. Reference numeral 2 denotes an A / D unit that digitizes an image signal obtained by the imaging unit 1 to obtain a digital signal of one bit or a plurality of bits for each pixel. Numeral 3 denotes a spatial differentiating means for performing a spatial primary differentiation process on the digitized image data and outputting a differential absolute value.

FIG. 2 is an explanatory diagram of the operation of the spatial differentiating means 3. As shown in the figure, when the luminance values of the pixels of the local block S in the image are represented by A to I, the luminance change in the vertical direction is represented by ΔV = (A + B + C)-(G + H + I) The luminance change in the direction is represented by ΔH = (A + D + G) − (C + F + I). Thus, the differential absolute value | e | becomes | e | = (ΔH ² + ΔV ² ) ^0.5 . The magnitude of the luminance change is calculated by the differential absolute value | e |
That is, the edge strength is determined.

The differential absolute value | e | obtained in this manner is for the central pixel E of the local block S. By scanning this local block S, the differential absolute value | e |
Is required.

Reference numeral 4 denotes a comparing unit that compares a differential absolute value output for each pixel from the spatial differentiating unit 3 with a preset setting value,
The image is converted into a binary image (edge image) as shown in FIG.

Reference numeral 5 denotes a current image memory for temporarily storing an edge image output from the comparing means 4.

Reference numeral 6 denotes a histogram measuring means which projects the edge image output from the current image memory 5 in the horizontal direction and the vertical direction as shown in FIG. 3. As shown in FIG. 3 (c), a main edge is extracted.

Reference numeral 7 denotes a detection mask creating means for creating a plurality of blocks including a main edge from the edge image and using the blocks as a detection mask.

FIG. 4 shows an example of a detection mask creating means. FIG. 4 (a) corresponds to the main edge image shown in FIG. 3 (c).
, A labeling process for giving a unique number to each pixel of a continuous edge is performed. Next, for each of the labels 〜, a block expanded by a predetermined image is used as a detection mask. For example, in the case of a label, the range shown by the bold line in FIG. 4B is the detection mask M for the label (shaded portion). Similarly, a detection mask is created for each label.

Next, another example of the detection mask creating means is shown in FIG. For the main edge image shown in FIG. 3 (c), for example, first, the leftmost main edge E _d (the portion exceeding the threshold existing at the left end of the vertical histogram in FIG. 3 (b)) is extracted, The length X ₁ in the X direction (a value corresponding to the number of pixels) is obtained. Then, to this major edge E _d, by measuring the histogram in the horizontal direction, obtaining the Y direction length Y ₁ (value corresponding to the number of pixels). In this way, the main the edge E _d, since the X-direction length X _1, Y-direction of the detection mask M of the main edge E _d a square of the length Y ₁ and adjacent meet two pieces' is there. This process is performed for all the other main edges. This detection mask creating means does not require labeling processing as compared with the above-described creating means, so that the processing becomes simple.

Reference numeral 8 denotes the image memory for storing a main edge image one frame before and a detection mask.

Reference numeral 9 denotes determination means for determining whether the vehicle has stopped. The exclusive OR is calculated for all pixels of the previous image and the current image for each detection mask, and the degree of coincidence is calculated. That is, in the detection mask M of FIG. 4B, the calculation is performed by setting the hatched portion corresponding to the label to “1” and the other portions to “0”. If edges of the same shape exist at the same position in the detection mask M, the number of pixels for which the calculation result is 0 increases. In the detection mask M, the degree of coincidence is determined based on the ratio of the pixel whose operation result is 0 to the total number of pixels of the detection mask M.

The above processing is performed for all the detection masks. Then, when the number of detection masks determined to be equal to or more than a predetermined number, it is determined that the vehicle is stopped.

According to the above embodiment, even if the edge images do not completely match, the determination is made based on the number of matching detection masks. Even if there is a mismatch in a part of the area, the stop of the vehicle can be reliably detected.

〔The invention's effect〕

As described above, according to the vehicle detection device of the present invention, the histogram of the edge image is obtained, and the main edge obtained by performing the threshold processing on each of the horizontal direction and the vertical direction is obtained. Create a detection mask for each main edge, compare the previous image and the current image for each detection mask,
When a predetermined number of the detection masks match, it is determined that the vehicle is stopped. Therefore, a part of the edge of the vehicle disappears due to the presence of an obstacle or the like in front of the vehicle. Even if this happens, a vehicle detection device capable of reliably detecting the stop of the vehicle can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a schematic diagram showing a spatial differentiating means according to the above, FIG. 3 is an explanatory diagram of main edge extraction according to the above, and FIG. FIG. 4 is an explanatory view showing one embodiment of creating a detection mask according to the embodiment. FIG. 5 is an explanatory view showing another embodiment of the production of a detection mask. DESCRIPTION OF SYMBOLS 1 ... Image pick-up means, 2 ... A / D means 3 ... Spatial differentiation means 4, ... Comparison means 5 ... Current image memory, 6 ... Histogram measurement means 7 ... Detection mask creation means 8 ... Previous image Memory, 9 ... determination means

Claims (3)

(57) [Claims] An imaging unit for capturing an image of a monitoring area to obtain an image signal; an A / D unit for digitizing the image signal obtained by the imaging unit; Spatial differentiation means for performing a differentiation process and outputting a differential absolute value, a comparing means for extracting an edge image by comparing the differential absolute value with a preset setting value, and obtaining a histogram of the edge image, horizontal direction,
Histogram measurement means for obtaining a main edge by performing threshold processing in each of the vertical directions, a previous image memory for storing a main edge of a previous image, and detection for creating a detection mask using the main edge A vehicle detection device comprising: a mask creation unit; and an exclusive OR of a main edge of a current image and a main edge of a previous image for each of the detection masks, and a determination unit for determining stop of the vehicle based on the result. 2. A vehicle detection apparatus according to claim 1, wherein said detection mask creating means labels the main edge and expands a predetermined pixel for each labeling as a detection mask. 3. A detection mask, wherein said detection mask creating means uses a detection mask as a quadrilateral having two adjacent lengths in both directions obtained by projecting said main edge in a vertical direction and a horizontal direction. The vehicle detection device according to claim 1, wherein