Method for improving object detectability by improving three-dimensional detection sensitivity
By deploying two cameras in the camera detection scene and using approximately orthogonal shooting directions, the problem of low sensitivity for acquisition of tiny objects is solved, and the detection ability and image clarity of the camera system for tiny objects is improved.
Patent Information
- Application Number
- CN202210581164.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-26
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2042-05-26
AI Technical Summary
When the camera detects tiny objects in an approximate planar scene, the ratio of the mount height to the detection distance is too small, resulting in too low sensitivity to the acquisition of tiny objects in the detected area, making it difficult to detect.
Two cameras are deployed on the same side outside the observation area, and the detected object is covered in an approximately orthogonal shooting direction, improving the sensitivity to collecting three-dimensional scale information of tiny objects. Specifically, the angle A between the shooting radius of the two cameras should be in the range of 30 to 120 degrees, with the optimal range being 85 to 95 degrees to achieve higher clarity.
By improving the three-dimensional information acquisition of long-distance tiny objects in the detection scene, the resolution of local close-up images of the imaging system is improved, and the detectability of tiny objects is significantly improved.
Smart Images

Figure CN115032708B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of imaging technology, and in particular to a method for improving the detectability of an object by enhancing the three-dimensional scale detection sensitivity. Background Art
[0002] With the development of society, the application fields of detection imaging technology are becoming wider and wider, and the application scenarios are increasing. In real scenes, due to the large number of small target scenes, it is very necessary to improve the detection rate of small target scenes and it has important application value.
[0003] The present invention studies the problem that when a camera detects tiny objects in a nearly planar scene, the tiny objects in the detected area have low acquisition sensitivity and are difficult to detect due to a small ratio between the camera's mounting height and the detection distance, and proposes a solution. Summary of the invention
[0004] In order to solve the technical problems raised by the background technology, the present invention provides a method for improving the detectability of objects by enhancing the three-dimensional scale detection sensitivity, so as to solve the problem that when a camera detects tiny objects in a nearly planar scene, the tiny objects in the detected area have too low acquisition sensitivity and are difficult to detect due to the small ratio of the mounting height to the detection distance. The method can improve the camera's acquisition sensitivity of the scale information of the three spatial dimensions of the tiny objects in the scene, thereby improving the detectability of the tiny objects in the scene.
[0005] In order to achieve the above object, the present invention adopts the following technical solutions:
[0006] A method for improving the detectability of objects by enhancing the three-dimensional scale detection sensitivity. The method is to deploy two cameras at different positions on the same side outside the observation area for tiny detected objects in a scene, and simultaneously cover the detected objects in the observation area with approximately orthogonal shooting directions, so as to realize the collection of three-dimensional scale information of the tiny objects.
[0007] Furthermore, the two cameras are camera one and camera two, the angle between the shooting radius of camera one and the detection surface of the detected object is a, the angle between the shooting radius of camera two and the detection surface of the detected object is b, and with respect to the detection surface of the detected object, the angle between the shooting radius of camera one and the shooting radius of camera two is A.
[0008] Furthermore, the approximately orthogonal shooting directions are specifically: for the detection surface of the detected object, the angle A between the shooting radius of camera one and the shooting radius of camera two is in the range of 30 to 120 degrees, and the closer to 90 degrees, the higher the clarity.
[0009] Furthermore, the optimal range of the angle A is 85 to 95 degrees.
[0010] Furthermore, the camera is any camera including a compound eye camera.
[0011] Compared with the prior art, the present invention has the following beneficial effects:
[0012] The method of the present invention utilizes two cameras set up at different positions and on the same side of the detection scene to detect the same area. By acquiring the images detected by the two cameras, the three-dimensional information of distant tiny objects in the detection scene is improved, the resolution of local close-up images in the camera system is made higher, and the effect of obtaining higher detectability is achieved. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 This is a layout diagram of Example 1 of a method for improving three-dimensional scale detection sensitivity of the present invention;
[0014] Figure 2 This is a layout diagram of Example 2 of a method for improving three-dimensional scale detection sensitivity of the present invention;
[0015] Figure 3 This is a layout diagram of Example 3 of a method for improving three-dimensional scale detection sensitivity of the present invention;
[0016] Figure 4 It is a layout diagram of an application embodiment of a method for improving three-dimensional scale detection sensitivity of the present invention.
[0017] Figure 5 is the shooting surface of the camera;
[0018] Figure 6 It is the angle diagram between the main detection surface of the camera and the shooting radius;
[0019] Figure 7 is the resolution map of the camera;
[0020] Figure 8 It is the actual resolution diagram of the camera. DETAILED DESCRIPTION
[0021] The specific implementation modes provided by the present invention are described in detail below with reference to the accompanying drawings.
[0022] like Figure 1-3 As shown, a method for improving the detectability of objects by enhancing the three-dimensional scale detection sensitivity is described. The method is to deploy two cameras at a certain distance on the same side outside the observation area for tiny detected objects in the scene, and simultaneously cover the detected objects in the observation area with approximately orthogonal shooting directions, so as to realize the collection of three-dimensional scale information of the tiny objects.
[0023] The two cameras are camera one and camera two. The angle between the shooting radius of camera one and the detection surface of the detected object is a, and the angle between the shooting radius of camera two and the detection surface of the detected object is b. With respect to the detection surface of the detected object, the angle between the shooting radius of camera one and the shooting radius of camera two is A.
[0024] The approximately orthogonal shooting directions are specifically: for the detection surface of the detected object, the angle A between the shooting radius of camera one and the shooting radius of camera two is in the range of 30 to 120 degrees, and the closer to 90 degrees, the higher the clarity.
[0025] The camera is any camera including a compound eye camera.
[0026] The present invention utilizes two cameras set up at different positions and on the same side of the detection scene to detect the same area. By analyzing and processing the image information contained in each pixel point of the images detected by the two cameras, the resolution of distant tiny objects in the detection scene is improved, and the resolution of local close-up images in the camera system is made higher, thereby achieving the effect of obtaining higher detectability.
[0027] Embodiment 1:
[0028] like Figure 1 As shown in the figure, when two cameras (Camera 1 and Camera 2) detect the same area, both cameras can cover any position in the area. When Camera 2 detects, the detected object is far away, and the angle b between the main detection surface of the object and the shooting radius is very small, resulting in a large actual resolution and low image clarity, that is, Camera 2 has low acquisition sensitivity to the object; when Camera 1 detects, the detected object is close, and the angle a between the main detection surface of the object and the shooting radius is very large, resulting in a small actual resolution and high image clarity, that is, Camera 1 has high acquisition sensitivity to the object.
[0029] In the method of the present invention, two cameras are arranged at a certain distance (the distance is determined according to the actual situation such as the size of the site) in the detection site, so that they detect the same area together. The two cameras realize approximately orthogonal shooting of the main detection targets (one or more) in the common detection area. Approximately orthogonal shooting means that the angle A between the shooting radii of the two cameras for the same detection object is approximately 90°, which is about 85°-92° in this embodiment. The images collected by the two cameras for the same detected object are analyzed and processed, and the image information of each pixel point of the images detected by the two cameras is collected to improve the resolution of small objects at a long distance in the detection scene, so that the resolution of local close-up images in the camera system is higher, and the effect of obtaining higher detectability is achieved. In this embodiment, the best effect is achieved when the angle A is about 85°-92°.
[0030] Embodiment 2:
[0031] like Figure 2 As shown, when two cameras (camera No. 1 and camera No. 2) detect the same area, the two cameras achieve approximately orthogonal shooting of the main detection targets (one or more) in the common detection area. Approximately orthogonal shooting means that the angle A between the shooting radii of the two cameras for the same detection object is about 50°-60° in this embodiment. The images collected by the two cameras for the same detected object are analyzed and processed. The effective information of the detected surface of the object in the processed image is greatly enriched compared to the image taken by a single camera. The rich image information greatly improves the clarity of the object in the camera system image. When the angle A is 50°-60°, the effect is slightly worse than that of Example 1.
[0032] Embodiment 3:
[0033] like Figure 3 As shown, when two cameras (camera No. 1 and camera No. 2) detect the same area, the two cameras achieve approximately orthogonal shooting of the main detection targets (one or more) in the common detection area. Approximately orthogonal shooting means that the angle A between the shooting radii of the two cameras for the same detection object is about 100°-120° in this embodiment. The images collected by the two cameras 2 for the same detected object are analyzed and processed. The effective information of the detected surface of the object in the processed image is greatly enriched compared to the image taken by a single camera. The rich image information greatly improves the clarity of the object in the camera system image. When the angle A is 100°-120°, the effect is slightly worse than that of Example 1.
[0034] in conclusion:
[0035] 1) When the angle A between the shooting radius of the two cameras for the same detection object is fixed, the angle A is not affected by the angle of the detection surface of the detection object. In other words, no matter how the detection surface of the detection object changes with the shooting angle a of camera 1 or the shooting angle b of camera 2, the clarity is not affected. (In traditional technology, the clarity is affected by the shooting angle a or b, see the principle description section below)
[0036] 2) According to the detection site, a reasonable distance and height between camera 1 and camera 2 are set, so that when the two cameras detect the same area, the two cameras can achieve approximately orthogonal shooting of the main detection targets (one or more) in the common detection area, that is, the angle A is close to 90°, thereby improving the clarity of the image.
[0037] Embodiment 4:
[0038] Airport runway foreign object FOD application example
[0039] According to the actual situation of the airport runway, by setting a reasonable distance and height between Camera 1 and Camera 2, when the two cameras detect the same area, the two cameras can achieve approximately orthogonal shooting of the main detection targets (one or more) in the common detection area, that is, the angle A is close to 90°, thereby improving the clarity of the image.
[0040] like Figure 4 As shown in the figure, when the detection area is a runway with a length of 3,000 meters and a width of 50 meters, the compound eye sensor is set up 100 meters away from the runway, the interval between the compound eye sensors is 200 meters, and the setting height is 6.5 meters. Except for the two towers at the edge of the runway, each of which only deploys one compound eye sensor, the remaining towers are equipped with two compound eye sensors.
[0041] Each runway area is photographed by two compound eye sensors. Due to the high definition of the images and the use of this method for image acquisition and processing, foreign objects such as FOD on the runway can be detected to ensure the safety of the aircraft on the runway.
[0042] The camera shooting principle involved in the present invention is explained as follows:
[0043] 1. Camera shooting surface
[0044] like Figure 5 As shown, the camera shooting plane refers to the abstract spherical surface formed at the same shooting distance within the camera field of view.
[0045] 2. Camera shooting radius
[0046] The distance between the camera and the subject.
[0047] 3. Angle between main detection surface and shooting radius
[0048] like Figure 6 As shown, in actual detection, the included angle between the main detected surface of the detected object and the shooting radius piece ranges from 0° to 90°.
[0049] 4. Resolution
[0050] Resolution refers to the minimum length that can be expressed by a pixel in a single dimension, usually expressed in millimeters per pixel (mm / p).
[0051] Modern cameras capture images by collecting light waves reflected / radiated by the scene through the lens, and converging and projecting these collected light waves onto the target surface of the photosensitive chip. The photosensitive chip converts the light signal into an electrical signal, which can be stored and transmitted, and when needed, it can be converted into a light signal on the screen to form an image.
[0052] The target surface of the photosensitive chip, as the receiving surface for receiving and collecting light waves, is composed of countless tiny pixels. The more pixels there are, the higher the clarity of the image. Each pixel corresponds to a pixel of the uncompressed original image.
[0053] The resolution usually refers to the length of the scene in the tangent direction near the tangent point of the camera's shooting surface corresponding to one pixel, that is, the length corresponding to the angle between the object's shooting surface and the shooting radius is 90°.
[0054] like Figure 7 As shown in the figure, the resolution is
[0055] The resolution varies due to factors such as the different lenses used by the camera, the different photosensitive chips used, and the different shooting distances. Generally, the greater the resolution, the worse the image quality, and vice versa.
[0056] Clarity refers to the number of pixels that make up a scene with a length of one meter in the image. The unit is pixel per meter (p / m), which is more commonly used internationally. The conversion between clarity and resolution is:
[0057]
[0058] 5. Actual resolution
[0059] In the actual captured images, since the ratio of the camera mounting height to the detection distance is too small, that is, the angle between the main detected surface of the object and the shooting radius is much smaller than 90°, this causes the actual distance contained in one pixel of the camera to increase a lot, that is, the actual resolution increases, so the effective information of the image cannot be clearly displayed.
[0060] like Figure 8 As shown in the figure, the resolution calculated in the normal way is h, but the actual detection surface is the plane where L is located, that is, the actual resolution, that is, the length expressed by one pixel is L. This situation will cause the detection image to be easily disturbed by noise and the detection target is not clear.
[0061] Generally speaking, the smaller the angle between the detected surface of the object and the radius, the greater the resolution, the lower the clarity, and the less clear the object is. It can also be said that the camera has low acquisition sensitivity for objects in this plane (detected surface) direction; the larger the angle between the detected surface of the object and the radius (maximum 90°), the smaller the resolution, the higher the clarity, and the clearer the object is. It can also be said that the camera has high acquisition sensitivity for objects in this plane (detected surface) direction.
[0062] 6. Solution of the present invention
[0063] The present invention arranges two cameras at a certain distance (the distance is determined according to the actual situation such as the size of the site) at the detection site, so that they detect the same area together. The two cameras achieve approximately orthogonal shooting of the main detection targets (one or more) in the common detection area. Approximately orthogonal shooting means that the angle between the shooting radii of the two cameras for the same detection object is approximately 90°. The present invention analyzes and processes the images collected by the two cameras for the same detected object, and the effective information of the detected surface of the object in the processed image is enriched to the maximum extent compared to the image taken by a single camera. The enrichment of the image information greatly improves the clarity of the object in the image of the camera system, improves the overall quality of the detection image of the camera system, and achieves the effect of obtaining higher detectability.
[0064] The above embodiments are implemented based on the technical solution of the present invention, and detailed implementation methods and specific operation processes are given, but the protection scope of the present invention is not limited to the above embodiments. The methods used in the above embodiments are conventional methods unless otherwise specified.
Claims
1. A method for improving the detectability of an object by increasing the sensitivity of three-dimensional detection. It is characterized in that The method is to deploy two cameras at different positions on the same side outside the observation area for a tiny object to be detected in the scene, and simultaneously cover the object to be detected in the observation area in approximately orthogonal shooting directions, so as to realize the collection of three-dimensional scale information of the tiny object; The two cameras are camera 1 and camera 2, the angle between the shooting radius of camera 1 and the detection surface of the detected object is a, the angle between the shooting radius of camera 2 and the detection surface of the detected object is b, and with respect to the detection surface of the detected object, the angle between the shooting radius of camera 1 and the shooting radius of camera 2 is A; The approximately orthogonal shooting directions are specifically: with respect to the detection surface of the detected object, the angle A between the shooting radius of the camera one and the shooting radius of the camera two is in the range of 30 to 120 degrees.
2. A method for improving the detectability of an object by improving the three-dimensional detection sensitivity according to claim 1, It is characterized in that The optimal range of angle A is 85 to 95 degrees.
3. The method of improving the detectability of an object by improving the three-dimensional detection sensitivity according to claim 1, It is characterized in that The camera is any camera including a compound eye camera.
Citation Information
Patent Citations
Device for acquiring depth image, calibrating method and measuring method therefor
CN103900532A
Telecentric stereo visual measurement apparatus and method for applying same to object three-dimensional morphology micron-order precision measurement
CN106500625A
Needle-sheet-shaped particle content detection method based on binocular camera image acquisition and recognition
CN110033516A
High-temperature forging visual inspection system with vibration isolation capability
CN113552136A
Rail fastener anomaly detection device based on binocular vision and laser speckle
CN207248770U
