TWI699708B - Container space calculation device - Google Patents

Abstract

The present invention discloses a container space calculation device, which includes an image capture device, an image converter, an image edge detector, a binary circuit and an image recognition processor. When goods are placed on the lower side of the container, and the goods are placed from the back of the container to the front in order, capture the overlapping images of the side and the lower side of the goods after overlapping, and based on the image position data of the lower side and the back The overlapping image is image warped to form a bird's-eye image corresponding to the lower side. After binarizing the image intensity of the lower side and the first edge of the cargo on the bird's-eye image to form the first enhanced edge, find the first enhanced edge corresponding to the front line, the back line, and the goods closest to the front line according to the horizontal angle, thereby Accurately calculate the occupancy rate of the goods relative to the container, while reducing the calculation cost.

Description

Container space calculation device

The present invention relates to a computing device, and particularly relates to a container space computing device.

In the transportation industry, a large number of goods travel to and from the world every day. Usually, this kind of cargo is transported in trailers or other cargo containers, which can be easily installed on different vehicles for transportation to the destination.

It is very important to monitor the condition of the cargo in the container. For example, being able to know whether a particular container is loaded with goods is important for a transportation company or container owner. One way to know whether a particular container is loaded with cargo is to manually open the container door and look inward. However, this can be troublesome because this method requires a person to be located in the container, and this can be a very time-consuming process, especially in the case of monitoring a large number of containers. The current method of calculating the utilization rate of containers can be ultrasonic, radar or optical time of flight (ToF, Time of Flight). Ultrasonic waves may misjudge materials such as cotton and metal goods when receiving reflected waves. , Different from image judgment. Ultrasonic sensors are often difficult to configure and calibrate, and usually require manual adjustment by the user during operation. In addition, the accuracy of optical time-of-flight ranging is relatively high, but the angle is small and the cost is expensive.

Therefore, the present invention aims to solve the above-mentioned problems and proposes a container space calculation device to solve the conventional problems.

The main purpose of the present invention is to provide a container space calculation device that uses image recognition to determine the occupancy rate of goods relative to the container, so as to avoid misjudgments and complicated correction procedures, and reduce costs.

In order to achieve the above objective, the present invention provides a container space calculation device, which includes an image capture device, an image converter, an image edge detector, a binarization circuit and an image recognition processor. The position of the image capture device corresponds to the position of a container. The container is a rectangular parallelepiped. The rectangular parallelepiped has a front side, a rear side, a left side, a right side, an upper side and a lower side. The lower side has a back line and a front line opposite to each other. , The lower side is vertically connected to the rear side by the back line, and the lower side is vertically connected to the front side by the front line. When goods are placed on the lower side, and the goods are placed sequentially from the back line to the front line, the image capture device captures the overlapping images of the back side and the lower side of the goods. The image converter is electrically connected to the image capturer and receives the overlapped image. The image converter performs image warping on the overlapped image according to the image position data of the lower side and the back line to form a bird's-eye view corresponding to the lower side. The image edge detector is electrically connected to the image converter, and receives the bird's-eye view image, and detects the bottom side and the first edge of the cargo on the bird's-eye view image. The binarization circuit is electrically connected to the image edge detector and receives the bird's-eye view image and the first edge. The binarization circuit binarizes the image intensity of the first edge to form a first enhanced edge. The image recognition processor is electrically connected to the binarization circuit, and receives the bird’s-eye view image and the first enhanced edge. The image recognition processor finds the first enhanced edge corresponding to the front line, the back line, and the goods closest to the front line according to the horizontal angle, thereby Calculate the occupancy rate of the goods relative to the container.

In an embodiment of the present invention, the image edge detector is electrically connected to the image capturer, the image recognition processor is electrically connected to the image converter, and the image position data includes the two rear end points of the back line and the two front end points of the front line For the position information, the lower side has two opposite side lines, and the two side lines vertically connect the back line and the front line. When the container is empty, the image capturer captures the image of the container with the back side and the bottom side. The image edge detector receives the container image and detects the second edge of the back side and the bottom side on the container image. The binarization circuit receives the container image and the second edge, and binarizes the image intensity of the second edge to form a second enhanced edge. The image recognition processor receives the container image and the second enhanced edge, and based on the orientation, horizontal angle, Two inclination angles and a first preset length are used to find the second enhanced edge corresponding to the back line and the two side lines. The image recognition processor uses the second enhanced edge corresponding to the back line to obtain the position information of the two back end points, and according to the second The preset length and the second enhanced edge corresponding to the two side lines obtain the position information of the two front end points, and the image recognition processor transmits the position information of the two rear end points and the two front end points to the image converter.

In an embodiment of the present invention, the binarization circuit uses Otsu's algorithm to binarize the image intensity of the second edge.

In an embodiment of the present invention, the image recognition processor uses Hough transform to find the second enhanced edge corresponding to the back line and the two side lines.

In one embodiment of the present invention, the image edge detector uses Sobel edge detection or Canny edge detection to detect the second edge.

In an embodiment of the present invention, the container space calculation device further includes an infrared light source electrically connected to the image capture device, and the image capture device cooperates with the infrared light emitted by the infrared light source to capture the overlapping image.

In one embodiment of the present invention, the container space calculation device further includes a wavelet converter, which is electrically connected to the image converter, the image edge detector, and the image capture device. The wavelet converter receives the overlapped image, and calculates the overlapped image. After the image undergoes two-dimensional wavelet transformation, the overlapping image is sent to the image converter for image distortion.

In an embodiment of the present invention, the binarization circuit uses Otsu's algorithm to binarize the image intensity of the first edge.

In an embodiment of the present invention, the image recognition processor uses Hough transform to find the first enhanced edge corresponding to the front line, the back line and the goods closest to the front line.

In an embodiment of the present invention, the image edge detector uses Sobel edge detection or Canny edge detection to detect the first edge.

In order to make your reviewer have a better understanding and understanding of the structural features of the present invention and the achieved effects, a preferred embodiment diagram and detailed description are provided. The description is as follows:

The embodiments of the present invention will be further explained by following relevant drawings. As far as possible, in the drawings and description, the same reference numerals represent the same or similar components. In the drawings, the shape and thickness may be exaggerated for simplicity and convenience. It can be understood that the elements not particularly shown in the drawings or described in the specification are in the form known to those skilled in the art. Those skilled in the art can make various changes and modifications based on the content of the present invention.

When an element is called "on", it can generally mean that the element is directly on other elements, or there can be other elements existing in both. Conversely, when a component is called "directly in" another component, it cannot have other components in between. As used herein, the term "and/or" includes any combination of one or more of the listed associated items.

The following description of "one embodiment" or "an embodiment" refers to at least one specific element, structure, or feature related to the embodiment. Therefore, multiple descriptions of "one embodiment" or "an embodiment" appearing in various places below are not directed to the same embodiment. Furthermore, specific components, structures, and features in one or more embodiments can be combined in an appropriate manner.

Please refer to Figure 1, Figure 2, Figure 3 and Figure 4 below, and introduce the first embodiment of the container space calculation device 10 of the present invention. The container space computing device 10 includes an image capturer 12, an image converter 14, an image edge detector 16, a binarization circuit 18, an image recognition processor 20, and a first wireless interface 22. The image capture device 12 is, for example, a digital camera, and its position corresponds to the position of a container 24. The container 24 is a cuboid. The cuboid has a front side 26, a rear side 28, a left side 30, a right side 32, an upper side 34 and a lower side 36. The lower side 36 has a rear line 38 and a front line 40 opposite to each other. In contrast to the two side lines 42, the lower side 36 and the back line 38 are vertically connected to the back side 28, the lower side 36 is vertically connected to the front line 40 to the front side 26, and the two side lines 42 are vertically connected to the back line 38 and the front line 40. The front side 26 has a top line 44 opposite to the front line 40. In addition, the container 24 has a known length L, width W, and height H, and the remaining structural dimensions of the container are also known. Therefore, the length of the rear line 38 and the front line 40 is equal to the length of the width W, and the length of the side line 42 is equal to the length. The length of L. For example, the image capture device 12 may be located on the top line 44, such as in the middle of the top line 44.

When the cargo 46 is placed on the lower side 36, and the cargo 46 is placed from the rear line 38 to the front line 40 in sequence, the lens of the image capture device 12 faces the rear line 38 and the position of the cargo 46 closest to the front line 40, and the image The extractor 12 captures the overlapping image S of the cargo 46 overlapping the rear side 28 and the lower side 36 so that the overlapping image S must include the position of the rear line 38 and the cargo 46 closest to the front line 40. The image converter 14 is electrically connected to the image capture device 12 and receives the overlapping image S. The image converter 14 performs image warping on the overlapping image S according to the image position data P of the lower side 36 and the back line 38 to A bird's-eye view image B corresponding to the lower side 36 is formed. The bird's-eye view image B must include the front line 40, the back line 38, the two side lines 42 and the position of the cargo 46 closest to the front line 40. The image edge detector 16 is electrically connected to the image converter 14 and receives the bird's-eye view image B, and detects the lower side 36 and the first edge E1 of the cargo 46 on the bird's-eye image B, such as the image edge detector 16 Sobel edge detection method or Canny edge detection method detects the first edge E1. Take Canny edge detection as an example, which mainly uses Gaussian filter to reduce image noise Then, the brightness gradient in the image is searched, and the threshold is used to classify whether each pixel belongs to an edge point to obtain the edge of the object, but the invention is not limited to this. The binarization circuit 18 is electrically connected to the image edge detector 16 and receives the bird's-eye view image B and the first edge E1. The binarization circuit 18 binarizes the image intensity of the first edge E1, leaving a stronger first edge E1 to form the first enhanced edge EH1. For example, the binarization circuit 18 uses Otsu's algorithm to binarize the image intensity of the first edge E1. The image recognition processor 20 is electrically connected to the binarization circuit 18, and receives the bird's-eye view image B and the first enhanced edge EH1. Because the lower side 36 undergoes a bird’s-eye view conversion, the first enhanced edge EH1 of the front line 40, the back line 38 and the cargo 46 will all be horizontal, so the image recognition processor 20 is based on the horizontal angle or the length of the horizontal angle, orientation, and width W. Find the first enhanced edge EH1 corresponding to the front line 40, the back line 38 and the cargo 46 closest to the front line 40 to calculate the occupancy rate R of the cargo 46 relative to the container 24. For example, the image recognition processor 20 uses Hough transform ) Find the first reinforced edge EH1 corresponding to the front line 40, the back line 38 and the cargo 46 closest to the front line 40, where the orientation can indicate the first reinforced edge EH1 of the front line 40, the back line 38 and the cargo 46 closest to the front line 40 relative position. Hough transformation is a method for recognizing geometric shapes in image processing. It is widely used in image processing and is not affected by the rotation of graphics. It is easy to perform rapid transformation of geometric graphics. The present invention is not limited to this. Specifically, since the scale of the bird’s-eye view image relative to the actual environment and the length of the side line 42 are known, the image recognition processor 20 matches this with the shortest distance between the front line 40 and the back line 38 in the bird’s-eye view image and the distance closest to the front line 40. The shortest distance between the first reinforced edge EH1 of the cargo 46 and the rear line 38 can accurately calculate the occupancy rate R of the cargo 46 relative to the container 24 to avoid misjudgment and complicated calibration procedures. At the same time, it can be identified by image To reduce costs. The first wireless interface 22 is electrically connected to the image recognition processor 20 and is wirelessly connected to a second wireless interface 48. The second wireless interface 48 is electrically connected to a control host 50. The first wireless interface 22 and the second wireless interface 48 are, for example, It is a Bluetooth transmission interface, but the invention is not limited to this.

The image position data P of the lower side 36 and the back line 38 can be preset or obtained in advance. The image position data P includes the position information of the two rear end points of the back line 38 and the two front end points of the front line 40. Please refer to Figure 1, Figure 5, Figure 6, and Figure 7. If the image position data P is obtained in advance, the image edge detector 16 is electrically connected to the image capturer 12, the image recognition processor 20 is electrically connected to the image converter 14, and the lens of the image capturer 12 faces the rear line 38 and The location of the second side line 42. When the container 46 is empty, the image capture device 12 captures the container image T with the rear side 28 and the lower side 36, so that the container image T must include the positions of the rear line 38 and the second side line 42. The image edge detector 16 Receive the container image T, and detect the second edge E2 of the rear side 28 and the lower side 36 on the container image T. For example, the image edge detector 16 uses Sobel edge detection or Canny The Canny edge detection method detects the second edge E2. The binarization circuit 18 receives the container image T and the second edge E2, and binarizes the image intensity of the second edge E2, leaving a stronger second edge E2 to form a second enhanced edge EH2, such as a binarization circuit 18Using Otsu's algorithm to binarize the image intensity of the second edge E2. The image recognition processor 20 receives the container image T and the second enhanced edge EH2. Since in the container image T, the back line 38 is horizontal, the length of the back line 38 is a first preset length, the lengths of the two side lines 42 are two second preset lengths, and the angles of the two side lines 42 are two obliques. Therefore, the image recognition processor 20 finds the second enhanced edge EH2 corresponding to the back line 38 and the two side lines 42 according to the known azimuth, horizontal angle, the two inclination angles, and the first predetermined length. For example, the image recognition processor 20 uses the Hough transform to find the second enhanced edge EH2 corresponding to the back line 38 and the two lateral lines 42, where the orientation can indicate the opposite of the second enhanced edge EH2 of the back line 38 and the two lateral lines 42 position. The image recognition processor 20 uses the second enhanced edge corresponding to the back line 38 to obtain the position information of the two back end points, and obtains the position of the two front end points according to the two second preset lengths and the second enhanced edge corresponding to the two side lines 42 Information, the image recognition processor 20 transmits the position information of the two rear end points and the two front end points to the image converter 14.

Please refer to Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, and Fig. 7 below, and introduce the operation process of the first embodiment of the container space calculating device 10 of the present invention. First, when the container 46 is empty, the image capture device 12 captures the container image T having the rear side 28 and the lower side 36 and transmits the container image T to the image edge detector 16. The image edge detector 16 detects the second edge E2 of the rear side 28 and the lower side 36 on the container image T, and then transmits the container image T and the second edge E2 to the binarization circuit 18. The binarization circuit 18 binarizes the image intensity of the second edge E2 to form a second enhanced edge EH2, and transmits the container image T and the second enhanced edge EH2 to the image recognition processor 20. The image recognition processor 20 searches for the second enhanced edge EH2 corresponding to the back line 38 and the two side lines 42 according to the orientation, horizontal angle, inclination angle and the first preset length, and uses the second enhanced edge corresponding to the back line 38 to obtain the second back end According to the second preset length and the second enhanced edge corresponding to the two side lines 42, the position information of the two front end points is obtained. The position information of the two rear end points of the back line 38 and the two front ends of the front line 40 form image position data P, and the image recognition processor 20 transmits the image position data P to the image converter 14.

Then, when the cargo 46 is placed on the lower side 36, and the cargo 46 is sequentially placed from the rear line 38 to the front line 40, the image capture device 12 captures the overlapping images S of the cargo 46 overlapping the rear side 28 and the lower side 36, and The superimposed image S is sent to the image converter 14. The image converter 14 performs image distortion on the overlapping image S according to the image position data P to form a bird's-eye view image B corresponding to the lower side 36 and transmits the bird's-eye view image B to the image edge detector 16. The image edge detector 16 detects the first edge E1 of the lower side 36 and the cargo 46 on the bird's-eye image B, and transmits the bird's-eye image B and the first edge E1 to the binarization circuit 18. The binarization circuit 18 binarizes the image intensity of the first edge E1 to form a first enhanced edge EH1, and transmits the bird's-eye view image B and the first enhanced edge EH1 to the image recognition processor 20. Finally, the image recognition processor 20 finds the first enhanced edge EH1 corresponding to the front line 40, the back line 38, and the cargo 46 closest to the front line 40 according to the horizontal angle or simultaneously with the horizontal angle, the orientation and the length of the width W, so as to calculate the cargo 46 relative to container 24 occupancy rate R. In addition, the image recognition processor 20 can further selectively transmit the occupancy rate R to the control host 50 through the first wireless interface 22 and the second wireless interface 48. Based on the travel route and occupancy rate R of the container 24, the control host 50 determines the feasibility of this travel route of the container 24 to fill the cargo with the remaining capacity. If the load cannot be completed, the control host 50 will notify other container users to go here The travel route is loaded with goods. If the goods can be loaded, the control host 50 does not need to notify users of other containers to go to this travel route to load the goods, so as to reduce the container budget and transportation costs.

Please refer to Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7 and Fig. 8 below, and introduce the second embodiment of the container space calculating device 10 of the present invention. The difference between the second embodiment and the first embodiment is that the second embodiment further includes an infrared light source 52 and a wavelet converter 54. The infrared light source 52 is electrically connected to the image capture device 12, and the wavelet converter 54 is electrically connected to the image converter. 14. Image edge detector 16 and image capture device 12. The image capture device 12 cooperates with the infrared light emitted by the infrared light source 52 to capture the overlapping image S or the container image T. However, because the container 46 is an indoor environment, in the absence of a light source, the image captured using infrared light has a weak signal in the container 46, so the image signal needs to be enhanced to make the edges of the image more robust. Therefore, compared with the first embodiment, the wavelet converter 54 in the second embodiment receives the overlapped image S, performs two-dimensional wavelet transformation on the overlapped image S, and transmits the overlapped image S to the image converter 14 for image distortion. . Alternatively, the wavelet converter 54 receives the container image T, performs a two-dimensional wavelet transformation on the container image T, and transmits the container image T to the image edge detector 16 to detect the second edge E2. Compared with the first embodiment, the related operations of the image processing, wireless transmission and control host 50 and the goals achieved in the second embodiment are the same as those of the first embodiment, and will not be repeated here.

In summary, the present invention uses image recognition to determine the occupancy rate of the goods relative to the container, so as to avoid misjudgments and complicated correction procedures, and reduce costs.

The above is only a preferred embodiment of the present invention, and is not used to limit the scope of implementation of the present invention. Therefore, all the shapes, structures, features and spirits described in the scope of the patent application of the present invention are equally changed and modified. , Should be included in the scope of patent application of the present invention.

10: Container space calculation device 12: Image grabber 14: Image converter 16: Image edge detector 18: Binary circuit 20: Image recognition processor 22: The first wireless interface 24: container 26: front side 28: rear side 30: left side 32: right side 34: upper side 36: lower side 38: back line 40: Frontline 42: Sideline 44: top line 46: Cargo 48: second wireless interface 50: control host 52: infrared light source 54: Wavelet converter

Figure 1 is a circuit block diagram of the first embodiment of the container space calculation device of the present invention. Figure 2 is a perspective view of an embodiment of the cargo container of the present invention. Figure 3 is a schematic diagram of the overlapping images of the present invention. Figure 4 is a schematic diagram of a bird's-eye view corresponding to the overlapping image of the present invention. Figure 5 is a perspective view of an embodiment of the empty cabinet of the present invention. Figure 6 is a schematic diagram of the container image of the present invention. Figure 7 is a schematic diagram of the bird's-eye view of the corresponding container image of the present invention. Figure 8 is a circuit block diagram of the second embodiment of the container space calculation device of the present invention.

10: Container space calculation device

12: Image grabber

14: Image converter

16: Image edge detector

18: Binary circuit

20: Image recognition processor

22: The first wireless interface

48: second wireless interface

50: control host

Claims (10)

A container space calculation device, comprising: an image capture device whose position corresponds to the position of a container. The container is a rectangular parallelepiped. The rectangular parallelepiped has a front side, a rear side, a left side, a right side, an upper side, and A lower side, the lower side has a rear line and a front line opposite to each other, the lower side is perpendicularly connected to the rear side by the rear line, the lower side is perpendicularly connected to the front side by the front line, goods are placed on the lower side, and the When the goods are placed sequentially from the back line to the front line, the image capture device captures the overlapping images of the goods overlapping the back side and the lower side; an image converter is electrically connected to the image capture device and receives For the overlapping image, the image converter performs image warping on the overlapping image according to the image position data of the lower side and the back line to form a bird's-eye view image corresponding to the lower side; an image edge detector, The image converter is electrically connected, and the bird's-eye view image is received, and the lower side and the first edge of the cargo are detected on the bird's-eye view image; a binary circuit is electrically connected to the image edge detector, And receiving the bird's-eye view image and the first edge, the binarization circuit binarizes the image intensity of the first edge to form a first enhanced edge; and an image recognition processor electrically connected to the binarization circuit, And receiving the bird’s-eye view image and the first enhanced edge, the image recognition processor finds the first enhanced edge corresponding to the front line, the back line and the cargo closest to the front line according to the horizontal angle, so as to calculate the relative relationship of the cargo The occupancy rate of the container. The container space computing device according to claim 1, wherein the image edge detector is electrically The image capture device is electrically connected, the image recognition processor is electrically connected to the image converter, the image position data includes the position information of the two rear end points of the back line and the two front end points of the front line, and the lower side has each other Relative to the two side lines, the two side lines vertically connect the back line and the front line. When the container is empty, the image capture device captures the image of the container with the back side and the bottom side, and the image edge detector Receive the container image, and detect the second edge of the back side and the lower side on the container image, the binarization circuit receives the container image and the second edge, and binarizes the second edge Image intensity to form a second enhanced edge. The image recognition processor receives the container image and the second enhanced edge, and searches for the corresponding back line and the second enhanced edge according to the orientation, horizontal angle, two tilt angles, and a first preset length For the second enhanced edge of the two side lines, the image recognition processor uses the second enhanced edge corresponding to the back line to obtain the position information of the two back end points, and according to the two second preset lengths and the corresponding two side lines The second enhanced edge obtains the position information of the two front end points, and the image recognition processor transmits the position information of the two rear end points and the two front end points to the image converter. The container space calculation device according to claim 2, wherein the binarization circuit uses Otsu's algorithm to binarize the image intensity of the second edge. The container space calculation device according to claim 2, wherein the image recognition processor uses Hough transform to find the second enhanced edge corresponding to the back line and the two side lines. The container space computing device according to claim 2, wherein the image edge detector detects the second edge by using Sobel edge detection or Canny edge detection . The container space calculation device according to claim 1, further comprising an infrared light source electrically connected to the image capturer, and the image capturer cooperates with the infrared light emitted by the infrared light source to capture the overlapping image. The container space calculation device according to claim 6, further comprising a wavelet converter, which is electrically connected to the image converter, the image edge detector and the image capture device, and the wavelet converter receives the overlapping image , And after performing a two-dimensional wavelet transformation on the overlapping image, the overlapping image is transmitted to the image converter to perform the image distortion. The container space calculation device according to claim 1, wherein the binarization circuit uses Otsu's algorithm to binarize the image intensity of the first edge. The container space calculation device according to claim 1, wherein the image recognition processor uses Hough transform to find the first enhanced edge corresponding to the front line, the back line and the cargo closest to the front line. The container space computing device according to claim 1, wherein the image edge detector detects the first edge by a Sobel edge detection method or a Canny edge detection method .

Images