KR101827114B1 - Apparatus and method for detecting proximal entity in pen - Google Patents

Abstract

In the detection of a near object in a housing, a proximity object region in which a plurality of objects are recognized as one is set as a region of interest in an image captured inside the housing through a camera section, and a second frame The second frame is projected on the first frame, and on the third frame according to the projection result, the morphological operation is performed on the object-by-object region of the second frame with respect to the region of interest Calculates the area ratio of the divided areas within the area of interest based on whether or not the area is overlapping, and adjusts the area of each object in the area of interest based on the area ratio to separate and display the individual objects.

Description

[0001] APPARATUS AND METHOD FOR DETECTING PROXIMAL ENTITY IN PEN [0002]

The present invention relates to an apparatus and a method for detecting an object within a housing based on a morphological operation process.

It is very difficult to grasp the behavior characteristics of livestock in real time due to the characteristics of barn and pig farms where a few administrators have to manage many objects. Recently, researches on object detection system that can construct low cost based on ICT (Information Communication Technology) have been actively carried out in small scale farms. Among them, object detection systems are being developed that continuously track and analyze livestock behavior through camera sensors.

In the past, object detection technology has been proposed to detect abnormal conditions in a house using various image processing techniques. Among them, the object detection method using the frame difference method calculates a pixel change between a previous frame (or a background frame) and a current frame and detects motion. In this differential image technique, the pixels in which the motion is detected are classified as moving objects, and numbers are assigned to the separated moving objects, thereby tracking the movement through them.

In this regard, Korean Patent Laid-Open No. 10-1998-031927 entitled "Surveillance camera and monitoring method for alarming the position of a moving object" has a lens state adjusting unit A storage unit for storing first and second image data for a surveillance region having a predetermined time interval, and a storage unit for storing a difference image between the first and second image data input from the storage unit, And a moving object detection unit for detecting a moving object depending on the size of the moving object and generating a moving object detection signal in accordance with the detected moving object, Lt; / RTI >

However, the object detection technique using this difference image technique is effective in detecting the behavior of each individual object on a frame basis, but there is a limitation in that the object overlapping problem can not be solved in continuous object tracking. In addition, there is an object detection technique using a moving picture camera as a method using a camera sensor other than the secondary image technique. In the case of such a moving image camera, there is a problem that it is affected by the light that varies depending on the weather and the time, and the characteristic information extracted during continuous monitoring of the object is distorted and is vulnerable to the problem of overlapping between the objects.

Accordingly, there is a need for an object detection technology that can separate multiple nearby objects by overcoming the above-mentioned problems, and can continuously detect and track motion of individual individual objects.

An embodiment of the present invention is to provide a device and method for detecting a housing proximity object which can separate an area recognized as a single object by a plurality of objects in proximity to each other in an image captured in a housing.

It should be understood, however, that the technical scope of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided an apparatus for detecting a close proximity to a housing, comprising: a camera unit for photographing the inside of the housing; A memory in which a proximity object detection program is stored; And a processor for executing a proximity object detection program stored in the memory, wherein, in response to the execution of the program, the processor extracts, from an image photographed through the camera unit, A morphological operation process is performed on a second frame which is a previous frame of the first frame in which the ROI is set and the morphological operation processed second frame is projected onto the first frame, Calculating a region ratio for the regions of interest within the region of interest based on whether the region of interest overlaps with the region of the object of the second frame on the third frame in accordance with a result of the comparison, The individual objects are separated and displayed by adjusting the divided regions for each object in the region of interest .

According to another aspect of the present invention, there is provided a proximity object detection method using a device for detecting a proximity of a close proximity to a building, the proximity object detection method comprising: step; Processing a morphological operation for a second frame that is a previous frame of the first frame in which the ROI is set; Projecting the morphologically processed second frame onto the first frame; Calculating a region ratio for the regions delimited within the region of interest based on whether the region of interest overlaps with the object-specific region of the second frame over the third frame in accordance with a result of the projection; And separating and displaying the individual entities by adjusting regions separated by the object within the region of interest based on the region ratios.

According to the above-mentioned object of the present invention, it is possible to accurately detect the overlapping region between individuals in a housing and pig house, which is a dense breeding environment, to distinguish and track individual objects. That is, even if two or more individuals in the detection image are recognized as a single entity, they can be individually separated. Therefore, it is possible to continuously manage and track individual objects individually.

FIG. 1 is a block diagram of an apparatus for detecting a nearby housing object according to an embodiment of the present invention.
2 is an example of a binarized image in which a proximity object is displayed according to an embodiment of the present invention.
3 is an example of an image obtained by labeling individual objects in an image before a plurality of entities approach each other according to an embodiment of the present invention.
FIG. 4 is a schematic diagram showing a result of an open calculation process for an image frame before generation of a proximity object according to an embodiment of the present invention.
FIG. 5 is a schematic diagram showing a result of erosion operation processing on an image frame before generation of a near object according to an embodiment of the present invention.
FIG. 6 is an example of an image frame obtained by projecting a morphological operation result frame on an image frame prior to proximity object generation according to an embodiment of the present invention to a current image frame.
FIG. 7 is a conceptual diagram showing regions of interest recognized as one object and separated into individual objects according to an embodiment of the present invention. Referring to FIG.
FIG. 8 is a flowchart illustrating a method for detecting a housing close-up object according to an exemplary embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description in the drawings are omitted, and like parts are denoted by similar reference numerals throughout the specification. In the following description with reference to the drawings, the same reference numerals will be used to designate the same names, and the reference numerals are merely for convenience of description, and the concepts, features, and functions Or the effect is not limited to interpretation.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when a component is referred to as "comprising ", it is understood that it may include other components as well as other components, But do not preclude the presence or addition of a feature, a number, a step, an operation, an element, a component, or a combination thereof.

Herein, the term " part " or " module " means a unit realized by hardware or software, a unit realized by using both, and a unit realized by using two or more hardware Or two or more units may be realized by one hardware.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an apparatus and method for detecting an in-housing proximity object according to an embodiment of the present invention will be described in detail with reference to the drawings.

1 is a block diagram of a housing detection apparatus according to an embodiment of the present invention.

1, a housing detection apparatus 100 includes a camera unit 110, a memory 120, and a processor 130. [

In an embodiment of the present invention, it will be described that the cage detection apparatus 100 detects and tracks a porcine entity being densely packed in a cage, considering a cage as an example of a cage. In addition, the pig body existing in the image photographed through the camera unit 110 is referred to as an object to be detected.

The camera unit 110 is installed at a position where the entire interior of the housing can be photographed at one position of the housing, and transmits the image photographed inside the housing to the processor 130 in real time. For example, as shown in FIG. 1, the camera unit 110 may be installed in a certain area (for example, a ceiling) at an arbitrary height from the floor.

The memory 120 separates the object by processing the housing photographed image, detects a proximity object in which a plurality of objects are approximated to be recognized as one object, separates the proximity object into individual objects, You are saving the program.

The processor 130 executes a program stored in the memory 120 to perform predetermined processes necessary for detecting a nearby entity. At this time, the processor 130 performs a morphological operation on a previous frame of the detected image frame (hereinafter, referred to as a "current frame") (ie, a frame before the object comes close). Then, the processor 130 projects the morphologically processed previous frame onto the current frame and adds the area enhancement algorithm and the adjacent pixel chrominance correction based on the area ratio to a frame obtained as a result of the projection (hereinafter, referred to as a 'projection result frame' Lt; / RTI > In this manner, the processor 130 separates the proximity object area, into which the plurality of objects have been recognized as one object, as individual objects. In this way, the processor 130 can utilize the information of the previous frame and the current frame at the same time by using the time information, so that individual objects separated from the previous frame can be applied to the area of the near object of the current frame.

Hereinafter, a processing procedure for detecting an object in the house through the processor 130 will be described in detail.

The processor 130 performs predetermined image processing on the image received from the camera unit 110 (that is, the image captured inside the housing) through the execution of the proximity object detection program, Pigs) and background.

At this time, the processor 130 sets a value of HSV (Hue Saturation Value) for the received image frame, and binarizes the resultant image to separate the object and the background. For reference, HSV is a color model that displays colors in the form of Hue, Saturation, and Value in graphics. Through such image processing, it is possible to remove noise substances or shapes (eg, object dirt or shadow of an object due to illumination) existing on the floor of the house from the original image of the house and distinguish the objects.

Then, the processor 130 detects a case where a plurality of objects among the detected objects are close to each other, and sets a region in which the proximity object exists as a region of interest (ROI). For example, the processor 130 may identify a proximity entity based on the detected area per object. At this time, the processor 130 may detect each object from the reference image frame that has been confirmed as having no nearby object in advance, and may store the area value of each object. Then, the processor 130 compares the reference area value set based on the area value of each object stored in advance with the area value of any object detected in the currently inputted frame, and if the area value of an arbitrary object exceeds the reference area value The object can be judged as a proximity object.

The processor 130 determines whether the current frame (i.e., the frame in which the proximity object has been generated) and the current frame before the interest region are generated (i.e., before the plurality of objects come close to each other ) Frames and stores them.

2 is an example of a binarized image in which a proximity object is displayed according to an embodiment of the present invention.

FIG. 2 shows an example of a region of interest P10, which is a partial region in which three swine individuals are close to each other in the entire region of the received image.

The processor 130 labels each object in a 'previous frame', which is a frame before a plurality of objects are brought close to each other, and identifies different objects so that they can be distinguished from each other. For reference, the labeling process refers to a process of labeling each object when the object and the background are separated from each other in the binary image. For example, when an image is binarized, a set of pixels having predetermined color values set in advance can be defined as one object.

FIG. 3 is an example of an image obtained by labeling an individual object in an image prior to object proximity according to an exemplary embodiment of the present invention.

In FIG. 3, the objects classified in the 'frame before the ROI is set' shown in FIG. 2 are shown. At this time, the region of interest is a result of the proximity of the three objects, and can be labeled as the first to third objects, respectively, as shown in FIG.

The processor 130 compares the pixel position of each object of the labeled previous frame with the object of interest in the current frame and if the position of the object in the two frames overlaps the predetermined area or more, can do.

On the other hand, the processor 130 processes the morphological operation for the previous frame before performing the projection processing.

The processor 130 generates a temporary frame (hereinafter referred to as a morphological operation temporary frame) by copying the previous frame matched and stored in the current frame. The processor 130 may perform morphological operation processing on the morphological operation temporary frame, which is not the previous frame original, as follows. However, it is also possible to morphologically process the original of the previous frame and project it on the current frame.

First, the processor 130 may use a morphological operation, which is one of the computer vision techniques, to uniformize the boundary width of the image information. This morphological operation is a technique of analyzing and processing the shape of an image, and it uses a predefined shape element to scan each pixel of the image and convert the image.

FIG. 4 is a schematic diagram showing a result of an open calculation process for an image frame before generation of a proximity object according to an embodiment of the present invention.

The processor 130 performs an opening operation of a morphological operation technique on a morphological operation temporary frame in order to solve the problem of individual pig separation more accurately by minimizing a problem that may arise in a future operation. Accordingly, the shapes of the individual objects (i.e., the first to third objects) on the morphological operation temporary frame as shown in Fig. 3 are transformed as shown in Fig. For reference, the open operation is a method of performing an erosion operation on an image followed by an expansion operation, and is mainly used to remove a detail area from an object such as noise. In this case, the erosion operation and the expansion operation refer to the process of examining each pixel of the morphological operation temporal frame and referring to the neighboring pixels of the examined pixel to convert the image. The processor 130 arranges the shapes of the individual objects using an open operation (one erosion operation and one expansion operation) in the morphological operation temporary frame.

Next, the processor 130 repeats the erosion operation process during the morphological operation until the central portion of each object becomes a certain size in the morphological operation temporary frame. For reference, the erosion operation reduces the size of the object and extends the background.

FIG. 5 is a schematic diagram showing a result of erosion operation processing on an image frame before generation of a near object according to an embodiment of the present invention.

The processor 130 repeatedly performs the erosion operation until the area of each object with respect to the result of performing the erosion operation is smaller than the size of one pig (i.e., an arbitrary individual object) do. For example, FIG. 5 shows that the erosion operation is performed up to a point of 1/3 area for each object. At this time, when one object is divided into a plurality of portions during the erosion operation, the processor 130 may reflect only the portion having the largest area as the erosion result.

Next, the processor 130 repeats the dilation operation process during the morphological operation until each object reaches the center portion of the other object in the morphological operation temporary frame in which the central portion of each object is obtained. For reference, the size of the object expands and the background decreases during the dilation operation.

Then, the processor 130 divides the image information (i.e., the near object region of the current frame) into a plurality of regions (i.e., individual objects) using the converted image frame and temporal information through the morphological operation . In order to separate a plurality of objects from each other in the area-based image information, the processor 130 uses a region growing technique to determine whether to add the inspected pixel as a neighboring area by examining neighboring pixels . At this time, the processor 130 processes the object separation by applying the result of the morphological operation, calculates the ratio occupied by each object based on the time information so that the object separation result can be completely trusted, re- Separate objects into individual objects.

Specifically, the processor 130 may generate a new frame by projecting a morphological operation temporary frame that has undergone the processes up to the erosion operation and the dilation operation processing, and may divide the proximity object region recognized as one object into individual objects Into several regions.

FIG. 6 is an example of an image frame obtained by projecting a morphological operation result frame on an image frame prior to proximity object generation according to an embodiment of the present invention to a current image frame.

FIG. 6 is a schematic diagram illustrating a result of projecting a temporary frame image subjected to a morphological operation on a labeled previous frame on a current frame image in which no region is distinguished, and shows the types of all regions that can be projected.

As shown in Fig. 6, the projection result frame is a result of overlapping the object-specific region on the morphological operation temporary frame and the proximate individual region of the current frame, and the confirmation region, the overlapped uncertain region, do. At this time, the 'determination area' indicates all areas having a color labeled in the previous frame, and the 'overlapped uncertain area' indicates all areas having a new color other than the color labeled in the previous frame. The 'uncertainty region without overlap' is a portion where there is no change in color even when a morphological operation temporal frame is projected on the current frame (in FIG. 6, for example, the white portion of the overlapped object is described as an example) Space is the area represented by the movement of objects.

The processor 130 then applies a region extension algorithm to the uncertainty-free region of the generated region in the new frame (i.e., the projection result frame).

Specifically, the processor 130 processes the pixel search from the starting point for the region generated in the projection result frame, and determines whether there is a " white " pixel in the adjacent pixel for each pixel. At this time, if white pixels exist in adjacent pixels, the process of filling white pixel portions with adjacent colors is repeated until all pixels are tested. On the other hand, if white pixels do not exist in the adjacent pixels from the starting point, the corresponding starting point is changed to the new starting point, and pixel searching and filling of the adjacent pixels with the white pixels are repeated.

Next, the processor 130 determines whether or not the area of the object to be labeled in the previous frame is the pixel of the labeled object, And calculates the ratio of the area (hereinafter, referred to as " first area ratio ").

&Quot; (1) "

As shown in Equation (1), the first region ratio is calculated by dividing the number of pixels of the confirmed region of each object identified in the proximity region of the current frame by the number of pixels of the object region labeled in the previous frame in the projection result frame applied up to the region extension algorithm can do.

Then, the processor 130 selects a color having a smaller ratio of the first area ratio among the colors causing the overlap, for each indeterminate area having overlapping among the areas generated in the projection result frame, and labels the corresponding area.

For example, referring to FIG. 6, the overlapped uncertainty region P61 overlaps the colors of the second object and the third object. That is, the overlapping color of the indeterminate area P61 where the labeling color of the second object and the labeling color of the third object are overlapped. At this time, the processor 130 calculates the first area ratio based on the number of pixels of the confirmed area P62 as the second object and the number of pixels of the second object area labeled in the previous frame in the proximity object area, The first area ratio is calculated based on the number of pixels of the confirmed area P63 as the third object and the number of pixels of the third object area labeled in the previous frame. Then, the processor 130 compares the calculated two first area ratios to select individual objects having smaller values, and labels the overlapping indeterminate areas P61 with the labeling colors of the selected individual objects. If the value of the first area ratio corresponding to the third object is smaller than the ratio of the first area corresponding to the second object in FIG. 6, the overlapped uncertainty area P61 can be labeled with the color of the third object have.

Then, the processor 130 calculates the number of pixels of the object region labeled in the previous frame and the number of pixels of the separated object in the proximity object region of the current frame, and calculates the area ratio (hereinafter referred to as " second area ratio " ). At this time, the separated objects in the proximity object area of the current frame include the newly labeled area based on the first area ratio.

&Quot; (2) "

As shown in Equation (2), the second region ratio can be calculated by dividing the number of pixels of the object region according to the result of newly labeling the current frame by the number of pixels of the object region labeled in the previous frame. In this case, the result of the new labeling process is the labeling result for the entire area including the newly labeled area and the original confirmed area for each of the classified objects in the proximity object area.

Next, the processor 130 determines whether or not the value of the second area ratio for each object is equal to or smaller than a predetermined value in the projection result frame in which the area expansion for the 'indeterminate area without overlap' and the labeling process for the ' And checks whether the threshold value is over or under to readjust the separated object-specific area within the proximity object area.

At this time, when there is an object whose ratio of the second area among the objects in the proximity object area exceeds the threshold value, the processor 130 has the lowest ratio among the objects in which the second area ratio is lower than the threshold value Select the object. The processor 130 then rescales the adjacent pixels of the selected object (i.e., the object with the lowest second area ratio).

The processor 130 determines the area of the individual object by repeating the color rebalancing process on the objects until the ratio of the second area of the objects in the proximity object area reaches the predetermined normal range.

FIG. 7 is a conceptual diagram showing regions of interest recognized as one object and separated into individual objects according to an embodiment of the present invention. Referring to FIG.

As shown in FIG. 7, as a result of processing described above with reference to FIG. 2 to FIG. 6, all the object-specific regions in the near object region are determined and a final result frame separated into individual objects can be obtained.

On the other hand, the processor 130 may process the tracking separately for a plurality of entities in the house that contain separate entities. For example, the processor 130 may calculate the Euclidean distance between the current location information and the previous location information for each entity, and may index the entity so as to correspond to the identification information at the previous location. Through this, it is possible to analyze and solve object overlapping (that is, proximity object) even in a situation where several objects move at the same time, and to simultaneously process tracking of individual objects within one image.

Hereinafter, referring to FIG. 8, a method of detecting a nearby object using the apparatus for detecting a nearby object according to an embodiment of the present invention will be described in detail.

In operation S801, a photographed image is acquired from a camera photographed inside the housing, and a proximity object in which a plurality of objects are recognized as one is determined in operation S802.

As described above, the procedure for determining whether or not a nearby object is generated can be processed through image binarization, which will be described below, and it is also possible to detect occurrence of a nearby object through a separate image processing.

In step S803, the HSV value is adjusted for the current frame, which is determined to have generated the proximity object area, to remove shadow, joy and the like.

That is, the current frame is binarized to separate the object and the background.

Next, the neighboring object region is designated as a region of interest for the binarized image frame (S804).

At this time, the current frame in which the region of interest is set and the frame before the current frame (i.e., the frame before the proximity entity is generated) are matched and temporarily stored. The previous frame is a state in which each object is labeled so that different objects can be identified.

Then, a previous frame is copied to generate a temporary frame for performing the morphological operation process, and a predetermined morphological operation process is performed on the generated temporary frame (S805).

At this time, as the morphological operation processing, the open operation, the erosion operation and the expansion operation are performed. This morphological operation process corresponds to the process described in Figs. 4 to 6 above.

Next, the morphological operation processed temporary frame is projected on the current frame, and the individual objects of the previous frame and the proximate object region of the current frame are overlapped (S806).

A plurality of regions are distinguished on a new frame (i.e., a projection result frame) that has processed the overlap, and it is determined whether there is an uncertain region having no overlap among the divided regions (S807).

6, depending on the result of overlapping the near object region of the current frame and the object region of the previous frame, the projection result frame may include one of the object regions of the previous frame among the proximity object regions of the current frame, "Confined area" in which overlapping occurs so that objects can be distinguished from each other, "overlapping indeterminate areas" in which two or more of the near object areas of the current frame and the object areas of the previous frame overlap, And an " overlap-free indeterminate area " in which overlapping with the object area of the previous frame of the proximity object area of the current frame does not occur.

As a result of the determination, if there is an indeterminate area having no overlapping, the area extension algorithm is applied to the area to perform labeling processing (S808).

Accordingly, the pixels of the indeterminate region without overlap are filled with the same color as the color of the most suitable object (i.e., adjacent pixels).

Next, in step S809, the first area ratio is calculated on the basis of the determined area for each object in the proximity object area of the current frame and the final area for each object labeled in the previous frame with respect to the projection result frame.

At this time, the determined area for each object in the proximity object area includes an area to which the area expansion algorithm is applied.

As a result of the judgment in the above-described step S807, if there is no indeterminate area without overlapping, the step S809 of calculating the first area ratio is executed immediately.

Then, it is determined whether there is an overlapped uncertainty area on the projected result frame (S810).

If there is an overlapping uncertainty region, the object having the lowest ratio of the first region among the object regions causing overlapping is selected and labeled so as to correspond to the object (S811).

Next, the second area ratio is calculated based on the area of the object labeled in the current frame and the area of the object labeled in the previous frame (S812).

At this time, the area of the object labeled in the current frame includes the confirmed area of the separated object in the proximity object area, the area filled by the area extension algorithm, and the newly labeled area based on the first area ratio.

If it is determined in step S810 that there is no overlapped uncertainty area, a step S812 of calculating the second area ratio is performed.

Then, in step S813, it is determined whether the ratio of the second area of the classified objects in the proximity object area is included in the normal range based on the preset threshold value, according to the result of calculating the second area ratio.

If there is an object area in which the ratio of the second area exceeds the threshold value, the color of the corresponding area is readjusted to the color of the object area in which the ratio of the second area is relatively low (S814).

That is, when there is an object whose ratio of the second area exceeds the threshold value among the objects in the proximity object area, the second area ratio is lower than the threshold value and the object having the lowest ratio is selected. And reselects the adjacent pixels of the selected object (i.e., the object having the lowest second area ratio).

In this case, steps S812 to S814 are repeated until the ratio of the second area of the classified objects in the proximity entity area is included in the normal range based on the predetermined threshold value.

If it is determined in step S813 that the ratio of the second area of the objects in the proximity object area is within the normal range based on the predetermined threshold value, (S815).

The above-described method of detecting a housing proximity object according to an embodiment of the present invention can also be implemented in the form of a recording medium including instructions executable by a computer such as a program module executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. The computer readable medium may also include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes any information delivery media, including computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transport mechanism.

It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

100: Near object detection device in house
110:
120: Memory
130: Processor

Claims (12)

A device for detecting proximity to a housing nearby,
A camera section for photographing the interior of the housing;
A memory in which a proximity object detection program is stored; And
And a processor for executing a proximity object detection program stored in the memory,
Wherein the processor sets, in response to the execution of the program, a proximity object area in which a plurality of objects are recognized as being close to one another in an image photographed through the camera section, A morphological operation process is performed on the second frame, a morphological operation-processed second frame is projected on the first frame, and on the third frame according to a result of the projection, Based on whether or not an individual object is overlapped with an object-specific region of the object region, and adjusting an area of each object in the region of interest based on the region ratio, Display,
A first region overlapping with one of the object-specific regions of the second frame in the region of interest, based on whether or not the region of interest overlaps with the region of the object of the second frame with respect to the region of interest on the third frame; And a second overlapping region, and a third overlapping region.
delete The method according to claim 1,
The processor comprising:
Applying a value of an adjacent pixel to the pixels in the third area by applying a preset area expansion algorithm,
Transforming the pixel value in the second area based on an area ratio between the first area and the object-specific area of the second frame,
When the ratio of the region of the second frame to the region of the object of the second frame is within a predetermined region ratio range, the region of the region of interest in the second region including the pixel values of the second region and the third region, Detecting device for close proximity of a house to be separated.
The method according to claim 1,
The processor comprising:
Wherein the processing unit performs the open operation, the erase operation, and the at least one expansion operation during the morphological operation for the temporary frame in which the second frame is copied, and projects the temporary frame to the first frame.
The method according to claim 1,
The processor
And a HSV (Hue Saturation Value) value for one frame of the photographed image, and binarizes the image according to a result of the setting to separate the object and the noise.
The method according to claim 1,
The processor comprising:
And performs a labeling process on the individual objects on the second frame.
A proximity object detection method using a proximity object detection device,
Setting a proximity object area in which a plurality of objects are proximate and recognized as one in an image captured inside the housing through a camera unit;
Processing a morphological operation for a second frame that is a previous frame of the first frame in which the ROI is set;
Projecting the morphologically processed second frame onto the first frame;
Calculating a region ratio for the regions delimited within the region of interest based on whether the region of interest overlaps with the object-specific region of the second frame over the third frame in accordance with a result of the projection; And
And separating and displaying the individual objects by adjusting regions separated by the object within the region of interest based on the region ratios,
The regions of interest,
A first region overlapping with one of the object-specific regions of the second frame in the region of interest, based on whether or not the region of interest overlaps with the region of the object of the second frame with respect to the region of interest on the third frame; A second region overlapped with the second region, and a third region that is not overlapped.
delete 8. The method of claim 7,
Wherein the step of separating and displaying the individual objects by adjusting the divided regions of the objects in the region of interest comprises:
Transforming a pixel value by applying a value of an adjacent pixel to the pixels in the third region by applying a predetermined region expansion algorithm;
Transforming pixel values in the second region based on an area ratio between the first region and an object-by-object region of the second frame; And
When the ratio of the region of the second frame to the region of the object of the second frame is within a predetermined region ratio range, the region of the region of interest in the second region including the pixel values of the second region and the third region, And separating and processing the housing close-up object.
8. The method of claim 7,
Wherein processing the morphological operation for the second frame comprises:
Processing the erase operation, one or more erosion operations, and one or more expansion operations during the morphological operation on the temporary frame from which the second frame was copied,
And projecting the temporal frame processed with the morphological operation to the first frame.
8. The method of claim 7,
The step of setting the proximity entity region as a region of interest comprises:
Setting a HSV (Hue Saturation Value) value for one frame of the photographed image;
Binarizing an image according to a result of the setting to separate the object and noise; And
Detecting an area in which the plurality of entities are adjacent to one another and recognizing the binarized image as a region of interest; and setting the binarized object as a region of interest.
8. The method of claim 7,
Prior to projecting the second frame to the first frame,
Further comprising: performing labeling processing on individual entities on the second frame.

Images