WO2021128553A1 - 目标检测方法、电路、视障辅助设备、电子设备和介质 - Google Patents

目标检测方法、电路、视障辅助设备、电子设备和介质 Download PDF

Info

Publication number
WO2021128553A1
WO2021128553A1 PCT/CN2020/076152 CN2020076152W WO2021128553A1 WO 2021128553 A1 WO2021128553 A1 WO 2021128553A1 CN 2020076152 W CN2020076152 W CN 2020076152W WO 2021128553 A1 WO2021128553 A1 WO 2021128553A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
target object
bounding box
image sensor
prompt
Prior art date
Application number
PCT/CN2020/076152
Other languages
English (en)
French (fr)
Inventor
蔡海蛟
冯歆鹏
周骥
Original Assignee
上海肇观电子科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海肇观电子科技有限公司 filed Critical 上海肇观电子科技有限公司
Priority to US16/834,957 priority Critical patent/US10838056B1/en
Publication of WO2021128553A1 publication Critical patent/WO2021128553A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/63—Control of cameras or camera modules by using electronic viewfinders
    • H04N23/631—Graphical user interfaces [GUI] specially adapted for controlling image capture or setting capture parameters
    • H04N23/632—Graphical user interfaces [GUI] specially adapted for controlling image capture or setting capture parameters for displaying or modifying preview images prior to image capturing, e.g. variety of image resolutions or capturing parameters
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
    • G06V10/235—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on user input or interaction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/61—Control of cameras or camera modules based on recognised objects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/63—Control of cameras or camera modules by using electronic viewfinders
    • H04N23/633—Control of cameras or camera modules by using electronic viewfinders for displaying additional information relating to control or operation of the camera
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/63—Control of cameras or camera modules by using electronic viewfinders
    • H04N23/633—Control of cameras or camera modules by using electronic viewfinders for displaying additional information relating to control or operation of the camera
    • H04N23/635—Region indicators; Field of view indicators
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
    • H04N23/60—Control of cameras or camera modules
    • H04N23/64—Computer-aided capture of images, e.g. transfer from script file into camera, check of taken image quality, advice or proposal for image composition or decision on when to take image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/02—Recognising information on displays, dials, clocks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07—Target detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition

Definitions

  • the present disclosure relates to the field of image processing, and in particular to a target detection method, an electronic circuit, a visually impaired auxiliary device, an electronic device, and a computer-readable storage medium.
  • a target detection method including: acquiring an image collected by an image sensor; detecting whether at least a part of a target object exists in the image; and responding to the target not being detected in the image At least a part of the object provides a prompt for adjusting the target object to be within the field of view of the image sensor.
  • an electronic circuit including: a circuit configured to perform the steps of the above-mentioned method.
  • a visually impaired auxiliary device including: an image sensor configured to collect an image; and an electronic circuit configured to detect whether there is at least a part of a target object in the image, wherein, The electronic circuit is further configured to provide a prompt for adjusting the target object to within the field of view of the image sensor in response to at least a part of the target object not being detected in the image.
  • an electronic device including: a processor; and a memory storing a program, the program including instructions that, when executed by the processor, cause the processor to execute the foregoing Methods.
  • a non-transitory computer-readable storage medium storing a program, the program including instructions that, when executed by a processor of an electronic device, cause the electronic device to execute the foregoing Methods.
  • Fig. 1 is a flowchart showing a target detection method according to an exemplary embodiment of the present disclosure.
  • FIG. 2A shows a schematic diagram of a bounding box containing a part of a target object detected in an image
  • FIG. 2B shows a schematic diagram of a bounding box containing the entire target object detected in an image.
  • Fig. 3 shows a schematic diagram of an exemplary process of detecting whether at least a part of a target object exists in an image.
  • Fig. 4 shows an exemplary diagram for judging whether it is a complete bounding box through a distance threshold.
  • Fig. 5A, Fig. 5B, Fig. 5C, Fig. 5D, Fig. 5E, Fig. 5F show example situations in which the edges of the bounding box and the edges of the image overlap.
  • Figure 6 shows an example of judging based on the distance range between the edge of the bounding box and the corresponding edge of the image.
  • 7A, 7B, and 7C show example situations where the distance between the edge of the bounding box and the corresponding edge of the image is not within a certain distance range.
  • 8A and 8B show an example of the positional relationship between the center position of the bounding box and the center area of the image.
  • 9A, 9B, and 9C show example situations of the relationship between the area of the bounding box and the area of the image.
  • Figure 10 shows an example situation where the text line is slanted.
  • Fig. 11A, Fig. 11B, Fig. 11C, Fig. 11D, Fig. 11E, and Fig. 11F show examples of the judging method of the pairwise combination.
  • Fig. 12A, Fig. 12B, and Fig. 12C show examples in which three kinds of judgment conditions are combined for judgment.
  • FIG. 12D shows an example of judgment based on the angle of the text line.
  • FIG. 13 is a flowchart showing a target detection method according to another exemplary embodiment of the present disclosure.
  • FIGS. 14A and 14B respectively show a structural block diagram of a stand, such as a desktop stand, in an unfolded state and a folded state for a visually impaired assist device according to an exemplary embodiment of the present disclosure.
  • FIG. 15 is a structural block diagram showing an electronic device according to an exemplary embodiment of the present disclosure.
  • first, second, etc. to describe various elements is not intended to limit the positional relationship, timing relationship, or importance relationship of these elements. Such terms are only used for Distinguish one element from another.
  • first element and the second element may refer to the same instance of the element, and in some cases, based on the description of the context, they may also refer to different instances.
  • FIG. 1 is a flowchart showing a target detection method according to an exemplary embodiment of the present disclosure.
  • the target detection method includes: step S101, acquiring an image collected by an image sensor; step S102, detecting whether there is at least a part of the target object in the image; step S103, responding When at least a part of the target object is not detected in the image, a prompt for adjusting the target object to be within the field of view of the image sensor is provided.
  • step S101 an image collected by an image sensor is acquired, so that it can subsequently be detected whether at least a part of a target object exists in the image.
  • the image sensor used to collect images can perform static or dynamic image collection. It can be an independent device (such as a camera, a video camera, a camera, etc.), or it can be included in various electronic devices (such as a mobile phone, Computers, personal digital assistants, assistive devices for the visually impaired, tablet computers, reading assistive devices, wearable devices, etc.).
  • the image collected by the image sensor may be a preview image, but not necessarily a photographed image.
  • the image sensor may be provided on a user's wearable device, auxiliary reading glasses, handheld devices, etc., so that the acquired image may be an image set on the wearable device or auxiliary reading glasses, for example.
  • the image of the target object collected by the sensor may be provided on a user's wearable device, auxiliary reading glasses, handheld devices, etc., so that the acquired image may be an image set on the wearable device or auxiliary reading glasses, for example.
  • the image of the target object collected by the sensor may be provided on a user's wearable device, auxiliary reading glasses, handheld devices, etc.
  • the target object may be an object placed somewhere, or may be an object held in the user's hand or supported by the user, or the like.
  • the target object may have a regular shape or an irregular shape.
  • the target object may be unfixed, for example, it may be moved by the user, or it may be moved by a mobile device used to move the target object, for example, the target object that can be moved may be: ID card, business card, passport, driver's license , Reading materials, tablet computers, mobile phones, etc.; even, the target object itself can move, such as a vehicle.
  • the target object may also be immovable, such as a TV installed on a wall, a fixed bulletin board, and so on.
  • the target object may include, for example, various forms of text, numbers, characters, symbols, pictures, and the like.
  • the acquired image may directly be an image acquired by an image sensor, or may be an image that has undergone certain or some pre-processing based on the image acquired by the image sensor.
  • the pre-processing may include, for example, de-noising. , Contrast enhancement, resolution processing, etc.
  • the image can be acquired from the image sensor in real time, or the image can be acquired some time after the image sensor has acquired the image.
  • the acquired image may be a pre-screened image, for example, after multiple collections, a clearer image is selected.
  • step S102 after the image is acquired, it is detected whether there is at least a part of the target object in the image.
  • a target detection algorithm (such as R-CNN, Fast R-CNN, RFCN, etc.) may be used to detect the bounding box of the target object.
  • the target detection algorithm can output parameters representing the position of the corresponding bounding box.
  • parameters may include information such as the center coordinates, width, and height of the bounding box.
  • the coordinates of each vertex of the bounding box can be obtained according to the obtained information such as the center coordinates, width, and height of the bounding box, so as to obtain the corresponding bounding box.
  • the parameters used to represent the position of the bounding box may also include other parameter combinations capable of representing position information, and are not limited to the examples cited here. In fact, you can choose which parameters to use to represent the position according to the specific needs of target detection.
  • the bounding box detection of the target object may also adopt other detection methods for determining the bounding box according to the position of the target object.
  • step S103 may be performed to provide a prompt for adjusting the target object to be within the field of view of the image sensor.
  • the present disclosure it is possible to first determine whether there is a situation where at least a part of the target object cannot be detected (for example, a situation where a bounding box cannot be detected) to automatically determine the image quality, and if it is determined that this situation exists , It will also automatically give prompts suitable for the user (for example, prompts for adjusting the target object to the field of view of the image sensor) to help users quickly adjust the relative position of the target object and the image sensor, and more Quickly obtain an image that meets the quality requirements (for example, whether the position and/or size of the target object in the image is appropriate), thereby effectively saving processing resources and greatly shortening the processing time.
  • prompts suitable for the user for example, prompts for adjusting the target object to the field of view of the image sensor
  • the target detection method and related circuits and equipment of the present disclosure can be used in visually impaired assist and wearable devices. It is difficult for users with visual impairment to visually judge the image quality or the wearable device worn by the user does not provide a display screen. Under the circumstance, by automatically helping the user to pre-judge the image quality, and further automatically prompting the user, so as to obtain better quality images, helping users to obtain the required services faster (such as reading services, guidance services, etc.).
  • providing a prompt for adjusting the target object to be within the field of view of the image sensor may include: in response to the image sensor being a fixed-position device, providing to move the target object to Prompting the target object to enter the field of view of the image sensor; and in response to the image sensor being a device with an adjustable position, providing for moving the target object and/or the image sensor to make the target object A prompt to enter the field of view of the image sensor.
  • the target object can be moved to allow the target object to enter
  • the prompt in the field of view of the image sensor facilitates user operations and improves user experience.
  • the image sensor is installed in a hand-held visually impaired assistive device or wearable visually impaired assistive device with a relatively flexible and adjustable position, that is, the position of the image sensor is relatively flexible and adjustable, and it can not only provide mobility
  • the target object is prompted to make the target object enter the field of view of the image sensor, and a prompt to move the image sensor so that the target object enters the field of view of the image sensor may also be provided to increase
  • the flexibility of user operations improves the user experience.
  • a prompt in addition to providing a prompt for adjusting the target object to within the field of view of the image sensor, a prompt can also be provided that at least a part of the target object cannot be detected.
  • the prompt that at least a part of the target object cannot be detected may include "the boundary cannot be detected. "Frame”, “The bounding box containing at least part of the target object cannot be detected", or the prompt of "Determine that there is no bounding box containing at least part of the target object in the image", etc.
  • providing a prompt to move the target object may include providing a prompt to move the target object away and/or offset
  • providing a prompt to move the image sensor may include providing a prompt to move the image sensor (for example, , Move the image sensor itself, or move the device where the image sensor is located to move the image sensor) to make it move away and/or shift.
  • the bounding box may not be detected because the target object is too close or relative to the position of the framing frame of the image sensor (or The field of view is somewhat biased, so the target object (and/or the image sensor) can be moved away and/or offset so that the target object can enter the field of view of the image sensor as much as possible.
  • the field of view of the image sensor is close to but different from the user’s field of view. In comparison, adjusting the target object to the image sensor’s field of view is relative to the adjustment The range of the target object to the corresponding user's field of view will make the target detection result more accurate.
  • the distance between the target object and the image sensor can be adjusted to enter the field of view of the view frame more, or the target can be adjusted
  • the direction of the object and/or the image sensor (the up, down, left, and right directions) can more enter the field of view of the viewfinder frame.
  • the distance of the target object may be preferably adjusted.
  • the method may further include step S104: in response to detecting at least a part of the target object in the image, determining that the at least part of the target object is related to the Whether the geometric relationship of the image satisfies a specific condition; and in response to the geometric relationship between the at least a portion of the target object and the image satisfying the specific condition, providing adjustment of the relative position between the target object and the image sensor , So that the target object is located at an appropriate position within the field of view of the image sensor.
  • the detected bounding box may be a bounding box containing a part of the target object, or a bounding box containing the entire target object.
  • FIG. 2A shows a bounding box 803 (incomplete bounding box) that contains a part of the target object 802 detected in the image 801
  • FIG. 2B shows the bounding box 803 (incomplete bounding box) detected in the image 801', which contains The bounding box 803' (complete bounding box) of the entire target object 802.
  • the bounding box of the target object detected in the image at least encloses the target object captured in the image (the image may contain the entire target object, or it may only contain a part of the target object), that is, the bounding box of the target object at least Enclose a real border formed by the actual outline of the target object in the image.
  • the bounding box 803 shown in FIG. 2A includes not only a part of the target object 802, but also the area outside the collected target object 802, the bounding box 803 can also be the same as that of the collected target object 802.
  • the real contours of a part match (approximately match), and the bounding box at this time is the smallest bounding box surrounding the real contour of the part of the target object 802.
  • the complete bounding box 803' shown in FIG. 2B includes not only the entire target object 802, but also the area outside the target object 802, the complete bounding box 803' may also be the same as the entire target object 802.
  • the real contours match (approximately match), and the complete bounding box at this time is the smallest bounding box surrounding the entire real contour of the target object 802.
  • the “bounding box” described herein can mean that at least a part of the target object is included in the bounding box, that is, it can be all or part of the target object; and the “complete bounding box” can mean that the bounding box contains at least a part of the target object. The entire target audience. If the bounding box is not detected, it belongs to “the bounding box containing at least a part of the target object cannot be detected” or “it is determined that there is no bounding box containing at least a part of the target object in the image”.
  • the detected at least part of the target object in response to detecting at least a part of the target object in the image, it may be further determined whether the detected at least part of the target object includes the entire target object, that is, whether it can be detected Complete target audience. For example, after detecting the bounding box of the target object (that is, it is determined that there is at least a part of the target object in the image), it is then checked whether the bounding box belongs to the complete bounding box, that is, whether the bounding box contains the entire target object, to determine whether it can Complete target object detected.
  • the "detecting whether at least a part of the target object exists in the image" in step S102 may include steps S1021 and S1022, for example.
  • step S1021 it is detected whether there is a bounding box containing at least a part of the target object in the image.
  • target detection algorithms such as R-CNN, Fast R-CNN, RFCN, etc.
  • R-CNN Random R-CNN
  • RFCN Radio Network Controller
  • target detection algorithms can be used to detect the bounding box of the target object.
  • the target detection algorithm After the target detection algorithm detects the target object on the image, it can output parameters representing the position of the bounding box.
  • parameters may include the center coordinates, width, and height of the bounding box. In this way, the coordinates of each vertex of the bounding box can be obtained according to the obtained information of the center coordinates, width and height of the bounding box.
  • the bounding box of the detected target object may be a rectangular box, as shown in FIG. 2A and FIG. 2B.
  • the rectangular frame can be determined by, for example, the positions of four vertices.
  • the bounding box of the detected target object may also be a shape other than a rectangle, such as a circle, a polygon, an ellipse, and other regular shapes, or various irregular shapes.
  • a shape other than a rectangle such as a circle, a polygon, an ellipse, and other regular shapes, or various irregular shapes.
  • the present disclosure does not impose restrictions on the shape of the bounding box.
  • step S1021 if in step S1021, the bounding box is not detected in the image (that is, it is determined that at least a part of the target object is not detected in the image), then go to step S103 and provide A prompt for adjusting the target object to be within the field of view of the image sensor. In addition, it can also provide a hint that the bounding box cannot be detected.
  • the prompt that the bounding box cannot be detected belongs to a kind of prompt that at least a part of the target object cannot be detected.
  • step S1021 the bounding box is detected in the image, then go to step S1022 to determine whether there is an overlap between the edge of the bounding box and the corresponding edge of the image. Whether the bounding box is a complete bounding box containing the entire target object, and then it is determined whether the complete target object can be detected.
  • step S105 If there is an overlap between the edge of the bounding box and the corresponding edge of the image ("Yes"), it can be determined that the detected bounding box is not a complete bounding box (that is, a complete target object cannot be detected). At this time, go to step S105 to provide a prompt that the complete bounding box cannot be detected.
  • the prompt that the complete bounding box cannot be detected belongs to a kind of prompt that the complete target object cannot be detected.
  • the overlap between the edge of the bounding box and the corresponding edge of the image it may be that there is an overlap between at least a part of at least one edge of the bounding box and at least a part of the corresponding edge of the image. That is, it is not necessary that the entire edge overlaps to be overlapped, but the overlap of part of the edge (not the entire edge overlap) also belongs to overlap. Therefore, not only the overlap of the entire edge means that the complete bounding box cannot be detected, but also the overlap of a part of the edge also means that the complete bounding box cannot be detected.
  • the bounding box of the detected target object after the bounding box of the detected target object is determined, it can be further determined whether the bounding box of the detected target object in the image is a complete bounding box, that is, whether the image contains a complete target object , In other words, whether the bounding box detected in the image can encompass the entire target object.
  • the target object such as ID cards and reading materials (books, magazines and other paper-medium reading materials, or e-books and other electronic reading materials) with a regular shape
  • the target object is a physical book on a paper medium
  • a complete bounding box 803' is detected in the image 801', and the complete bounding box 803' surrounds each actual edge of the target object 802.
  • similar detection can also be performed, that is, to determine whether the detected bounding box in the image contains the complete target object, for example, it can be determined by whether the detected bounding box contains the outline of the target object.
  • step S1022 it can be determined whether the detected bounding box is a complete bounding box containing the entire target object by, for example, detecting whether there is an overlap between the edge of the bounding box and the corresponding edge of the image, and then determining whether the detected bounding box is a complete bounding box containing the entire target object. Whether at least a part of the detected target object includes the entire target object.
  • a prompt that the complete bounding box cannot be detected may be provided in step S105.
  • step S104 If there is no coincidence between the edge of the bounding box and the corresponding edge of the image (indicating that a complete bounding box is detected), then go to step S104.
  • the detected bounding box is a rectangular frame
  • Figures 5A and 5B are the cases where the edges of the detected bounding box and the corresponding edges of the image completely overlap
  • Figures 5C, 5D, 5E, and 5F are the edges of the detected bounding box and the image. The situation where the corresponding edges of are overlapped.
  • providing the prompt that the complete bounding box cannot be detected includes prompting to adjust the relative position between the target object and the image sensor, so that the target object is far away and/or shifted.
  • prompting to move the target object away and/or offset includes, for example, moving the target object away (zoom out) and/or offset, for example, it can be zoomed out and/or offset relative to the view frame of the image sensor.
  • move the image sensor for example, move the image sensor itself, or move the device where the image sensor is located, such as wearable devices or visually impaired reading devices to move the image sensor
  • the detected rectangular frame overlaps the image with four sides, it means that a part of the target object exceeds the field of view of the viewfinder frame, and the target object is not collected completely; or as shown in FIG. 5B, It means that the target object is just in the field of view of the viewfinder frame and it is fully captured, but the target object is too large in the image. For this situation where the four sides overlap, you can directly prompt to move the target object away, so that the target object enters the viewfinder more The field of view of the box.
  • the target object may not be collected completely, and you can directly prompt the target object Far away (for example, three edges coincide with each other in FIG. 5C, and two edges overlap in FIG. 5D), and/or move the target object in a direction related to the following orientation: the orientation passes through the bounding box, and The side of the image is determined by the direction opposite to the side where the side overlaps (for example, in Figure 5E, the two sides of the bounding box and the two sides of the image are overlapped in the upper left direction, respectively, and you can follow the lower right direction opposite to the upper left direction.
  • the complete bounding box is not detected.
  • the detected bounding box completely overlaps with the frame of the image, it can also be regarded as detecting a complete bounding box, that is, the present disclosure is not limited to regard this situation as not detecting a complete bounding box. , But you can decide whether to use the bounding box that completely coincides with the border of the image as a complete bounding box according to actual needs.
  • the bounding box of the target object and the border of the image completely overlap, which may affect the image recognition effect near the edge of the image, because the definition of the edge of the image collected by the current image sensor is relatively low. Therefore, when the collected target object is close to the center of the image, the recognition effect will be better. Therefore, by setting a specific distance threshold, the obtained bounding box can be made smaller than the frame of the image, so that the target object in the obtained bounding box can be completely and centered in the image as much as possible.
  • step S103 after step S103 provides a prompt to adjust the target object to be within the field of view of the image sensor, it is possible to go to step S101 to reacquire the image re-acquired by the image sensor. , And continue to detect whether there is at least a part of the target object in the reacquired image in step S102 (more specifically, step S1021 in FIG. 3), until at least a part of the target object is detected in the reacquired image.
  • prompts will be automatically provided to the user to help The user quickly adjusts the relative position of the target object and the image sensor, and obtains images that meet the quality requirements (such as whether the position and/or size of the target object in the image are appropriate) faster, thereby effectively saving processing resources and greatly Shorten the processing time.
  • the target detection method and related circuits and devices of the present disclosure can be used for visually impaired assistance and wearable devices.
  • users with visual impairments or users who wear wearable devices that do not provide a display screen cannot judge images.
  • the exemplary embodiments of the present disclosure can automatically help the user to pre-judge the image quality, and further automatically prompt the user to obtain a better quality image, so as to improve the accuracy and speed of target recognition, thereby effectively saving It saves processing resources, greatly shortens processing time, and helps users to obtain required services (such as reading services, guidance services, etc.) faster.
  • the providing the prompt that the bounding box is not detected includes prompting to move the target object away and/or offset.
  • the bounding box may not be detected because the target object is not within the acquisition range (ie, the field of view) of the image sensor.
  • the target object is moved away (away from the image sensor) (by moving the target object and/or moving the image).
  • the sensor can expand the collection range, and it is possible to accommodate the target object into the expanded collection range.
  • the position of the target object can be offset to make it enter the acquisition range of the image sensor.
  • the user can move the target object away from the image sensor to expand the acquisition range and at the same time offset the target object to enter the acquisition range of the image sensor.
  • step S104 if at least a part of the target object is detected in the image in step S102, then it can go to step S104.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, a prompt for adjusting the relative position between the target object and the image sensor is provided.
  • the "geometric relationship” here includes, but is not limited to, distance relationship, area ratio relationship, position relationship, and so on.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, providing a prompt for adjusting the relative position between the target object and the image sensor includes :
  • a prompt for adjusting the relative position between the target object and the image sensor is provided.
  • the specific distance requirement may include that the distance between each edge of the bounding box and the corresponding edge of the image is greater than a corresponding specific distance threshold.
  • the rectangular bounding box of the target object and the four sides of the image can respectively correspond to four specific distance thresholds, for example, they can be called upper distance threshold D up , lower distance threshold D down , and left distance threshold. D left , the right distance threshold D right (as shown by the dashed arrow in Figure 4).
  • the bounding box is a complete bounding box.
  • each distance threshold can be set according to actual needs, and they are not necessarily equal.
  • each distance threshold may be zero or a value greater than zero.
  • the detected bounding box is a complete bounding box, and the complete bounding box is smaller than the bounding box of the image.
  • each distance threshold can be set according to actual needs, and the present disclosure does not require additional restrictions on it.
  • each specific distance threshold may be set such that the obtained complete bounding box is smaller than the image frame, that is, there is no need to set the specific distance threshold here too large.
  • step S102 when at least one distance between the detected edge of the bounding box and the corresponding edge of the image is equal to 0, the situation that the edge of the bounding box described above overlaps with the corresponding edge of the image occurs, that is, When detecting the bounding box in step S102 (more specifically, step S1022) described above, it is determined whether there is an overlap between the edge of the bounding box and the edge of the image.
  • the tips include:
  • the target object moves in an orientation that is the opposite orientation of the center position of the bounding box relative to the center position or the center area of the image.
  • the specific distance requirement may further include that the distance between the edge of the bounding box and the corresponding edge of the image is within a corresponding specific distance range.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is possible to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • a prompt for adjusting the relative position between the target object and the image sensor is provided.
  • step S104 also uses the distance between the edge of the bounding box of the target object and the corresponding edge of the image
  • the previous embodiment detects the distance between each edge of the bounding box of the target object and the image. Whether the distance between the corresponding edges is greater than the corresponding specific distance threshold is used to determine whether to provide a prompt for adjusting the relative position of the target object and the image sensor.
  • the latter embodiment determines whether to provide adjustment of the relative position of the target object and the image sensor based on whether the distance between the edge of the bounding box and the corresponding edge of the image is within a corresponding specific distance range. prompt.
  • step S104 it is possible to determine whether the position of the detected bounding box is appropriate through the corresponding specific distance threshold and/or the corresponding specific distance range, and if it is not appropriate, for example, the bounding box may be determined Whether the position of is up, down, left, right, etc. (for example, less than the corresponding specific distance threshold, or less than the lower limit of the corresponding specific distance range or greater than the upper limit of the corresponding specific distance range, etc.).
  • the aforementioned judgment related to the specific distance threshold and the aforementioned judgment related to the specific distance range may be combined into a single judgment step for implementation. In other words, it can only be judged whether the distance between each edge of the detected bounding box and the corresponding edge of the image is within the corresponding specific distance range, or both the specific distance threshold judgment and the specific distance range judgment are performed. As long as the lower limit of the specific distance range is the specific distance threshold. On the other hand, when the lower limit of the specific distance range is greater than the specific distance threshold, it can be judged separately as described above. In short, the various embodiments of the present disclosure can be flexibly changed, and whether to judge separately here can depend on actual requirements and/or considerations of processing resources, and so on.
  • both the distance threshold and the distance range belong to the requirements for distance. Therefore, the “specific distance requirement” can be uniformly used in the present disclosure to express.
  • specific distance requirements the present disclosure includes but is not limited to the distance thresholds and distance ranges exemplified above.
  • the bounding box that can meet the relevant distance requirements is a complete bounding box.
  • the bounding box may be judged first, and then other conditions may be judged and/or position adjustment prompts may be made to obtain an image with a suitable position and/or size of the target object.
  • Figure 6 will specifically give an example of judging the distance range.
  • the detected bounding box is a rectangular box, for example, four specific distance ranges can be set, which are respectively referred to as the upper distance range D up-range , the lower distance range D down-range , The left distance range D left-range , the right distance range D right-range .
  • each distance range is defined by its respective lower limit of the distance range and upper limit of the distance range, and the aforementioned specific distance threshold is a value.
  • the lower limit of each distance range can be greater than or equal to the above-mentioned corresponding distance threshold.
  • the judgment condition of the specific distance range is relatively stricter than the judgment condition of the specific distance threshold.
  • the upper, lower, left, and right distance ranges are respectively shown by the dashed arrows in the figure, and the lower limit of the upper, lower, left, and right distance ranges constitute the large in Figure 6
  • the dashed frame, the upper limit of the upper, lower, left, and right distance ranges form the small dashed frame in FIG. 6 (for clarity, the target object in the bounding box is not shown in FIG. 6). If each side of the detected bounding box is located between the large dashed box and the small dashed box, the bounding box 803 in FIG.
  • each of the bounding boxes is The distance between the edge and the corresponding edge of the image is within the respective specific distance range (the distance between each edge of the bounding box and the corresponding edge of the image is greater than the lower limit of the respective distance range and less than the respective distance range The upper limit of ), it may not be necessary to prompt to adjust the relative position between the target object and the image sensor, or other conditions may be used to further determine whether it is necessary to prompt to adjust the relative position.
  • judging whether the distance between the edge of the bounding box and the corresponding edge of the image is within a certain distance range, that is, both the upper limit control for the distance between the edges and the
  • the lower limit of the distance between the edges is controlled to determine whether the position and/or size of the target object in the image is appropriate. For example, judging whether the distance between the edges is within a certain distance range can further reflect the image quality. For example, when a certain or some distances between the edges are less than the lower limit of the corresponding distance range, the target The object may be too close to the edge of the image.
  • the user can be prompted how to adjust the relative distance between the target object and the image sensor (to be described later) to make The target object leaves the edge of the viewfinder frame and is close to the center of the viewfinder frame, so that a better quality image can be obtained.
  • the position of the target object in the image may be offset or the size of the entire target object in the image is too small, which affects the image quality.
  • step S104 if the distance between the edge of the bounding box and the corresponding edge of the image is not within the corresponding specific distance range, it is provided to adjust the relationship between the target object and the image sensor.
  • the hints of the relative position between the two can include:
  • the distances between the upper, lower, left, and right sides of the bounding box and the corresponding upper, lower, left, and right sides of the image are respectively detected, and the upper distance between the bounding box and the image
  • the distance from the bottom, the distance from the left, and the distance from the right are the upper distance as the distance between the upper edge of the bounding box and the upper edge of the image, and the lower side as the distance between the lower edge of the bounding box and the lower edge of the image Distance, the left distance as the distance between the left side of the bounding box and the left side of the image, and the right distance as the distance between the right side of the bounding box and the right side of the image (due to space limitations, not shown in Figure 6 Out of these four distances); and
  • each of the distance ranges is determined by its respective upper limit and distance
  • the lower limit is jointly defined, if at least one of the upper distance, lower distance, left distance, and right distance is not within the respective upper distance range, lower distance range, left distance range, and right distance range. (That is, at least one of the upper distance, the lower distance, the left distance, and the right distance is not in the position between the large dotted frame and the small dotted frame shown in FIG. 6), then it is provided to adjust the target object and the A hint of the relative position between the image sensors.
  • the above-mentioned specific distance range can be set according to actual needs (as shown in Figure 6, the lower limit of the upper, lower, left, and right distance ranges and the upper, lower, left, and right distances The upper limit of the range), the present disclosure does not need to limit this.
  • the lower limit of the distance range can be 0 or a value greater than 0, and the upper limit of the distance range should be a value greater than the corresponding lower limit of the distance range.
  • the specific distance ranges on the upper side, the lower side, the left side, and the right side may not be equal. Of course, as in the example here, the four specific distance ranges are all equal.
  • FIG. 6 is only one for explaining how to determine whether to prompt to adjust the target object based on whether the distance between the edge of the bounding box and the corresponding edge of the image is within the corresponding specific distance range.
  • An example of the relative position between the image sensors Those skilled in the art will know that the present disclosure is not limited to this example.
  • step S104 in response to the distance between the edge of the bounding box and the corresponding edge of the image does not meet a specific distance requirement, it is provided to adjust the relative position between the target object and the image sensor
  • the tips include:
  • the edge of the bounding box and the corresponding edge of the image is greater than the upper limit of the corresponding specific distance range, it is prompted to bring the target object closer to and/or the edge of the bounding box greater than the corresponding edge.
  • the sides of the upper limit of the specific distance range are located in the same direction to offset the target object.
  • the user is prompted how to adjust the target object and the image.
  • the relative distance between the sensors makes the position and/or size of the target object in the image more suitable, so that a better quality image can be obtained.
  • Figures 7A, 7B, and 7C give several example situations where "the distance between the edge of the detected bounding box and the corresponding edge of the image is not within the corresponding specific distance range" for easy understanding and does not represent This disclosure is limited to these few example cases.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, the adjustment of the relative relationship between the target object and the image sensor is provided.
  • Location hints can include:
  • a prompt for adjusting the relative position between the target object and the image sensor is provided.
  • the center position of the bounding box may be determined by the position information of the bounding box in the image. Specifically, for example, the center position of the bounding box may be determined by the position of each vertex of the bounding box in the image. In addition to the vertex position, the center position of the bounding box can also be determined by other position information of the bounding box in the image, which is not specifically limited in the present disclosure.
  • the central area of the image may be an area surrounding the central point of the image.
  • the size and shape of the central area can be determined according to actual conditions, and the present disclosure does not impose special restrictions on its size and shape.
  • the shape of the central area for example, it may be a rectangle, a circle, or the like.
  • the center point here may be, for example, the geometric center. Therefore, the center point of the image may also be, for example, the center point (geometric center) of the central area.
  • Deviation from the center area (or center position) of the image will affect the quality of the image, and even further affect the services provided to users (such as visually impaired reading services, guidance services, etc.), which can also prompt the user how to adjust the target object and the image sensor The relative distance between them (which will be described later) to make the position of the target object in the image more suitable (such as centering as much as possible), so that a better quality image can be obtained.
  • FIG. 8A shows a situation where the center position 805 of the bounding box 803 of the detected target object is not in the center area 804 of the image 801. At this time, it is possible to adjust the relative position between the target object and the image sensor. Tips.
  • FIG. 8B shows a situation where the center position 805 of the bounding box 803 of the detected target object is in the center area 804 of the image 801. At this time, the next operation can be performed as needed.
  • providing a prompt for adjusting the relative position between the target object and the image sensor may include:
  • the user is prompted to move the target object in a direction opposite to the direction.
  • the detected center position 805 of the bounding box 803 is not in the center area 804 of the image 801. It can be seen from FIG. 8A that the center position 805 is in the upper right direction of the central area 804, it can be determined that the target object can be moved in the lower left direction opposite to the upper right direction, so that the center position 805 of the bounding box 803 enters the central area 804 of the image 801 in.
  • the user is prompted to adjust the relative distance between the target object and the image sensor. , So that the target object is more suitable in the image, so that a better quality image can be obtained.
  • the response in step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfies a specific condition, providing adjustments between the target object and the image sensor Tips for the relative position of can include:
  • a prompt for adjusting the relative position between the target object and the image sensor is provided.
  • the area ratio is judged to determine whether the size of the target object collected in the image is appropriate.
  • the ratio of the area of the bounding box to the area of the image is about 1:9, which is too small; as shown in Figure 9B, the ratio of the area of the bounding box to the area of the image is about 94:100, the ratio Too large; as shown in Figure 9C, the ratio of the area of the bounding box to the area of the image is approximately 3:4, and the area ratio is appropriate.
  • the target object in the image is appropriate by judging whether the ratio of the area of the bounding box to the area of the image is within a specific ratio range. For example, judging whether the ratio of the area of the bounding box to the area of the image is in a specific ratio range (the image that fits the ratio range has a more appropriate image size and better image quality), which can further reflect the image quality, such as when the area of the bounding box
  • the ratio to the area of the image is not in a specific ratio range, the target object may be too large or too small in the image, which affects the quality of the image and even further affects the services provided to users (such as visually impaired reading services, guidance services, etc.), Therefore, it is also possible to prompt the user how to adjust the relative distance between the target object and the image sensor (to be described later), so that the size of the target object in the image is more suitable, so that a better quality image can be obtained.
  • step S104 if the ratio of the area of the bounding box to the area of the image is not within a specific ratio range, a method for adjusting the relative position between the target object and the image sensor is provided. Tips can include:
  • the area ratio is too small as shown in FIG. 9A, it may be prompted to bring the target object closer. Conversely, for the case where the area ratio is too large as shown in FIG. 9B, it can be prompted to keep the target object away.
  • the user is prompted how to adjust the target object and the image.
  • the relative distance between the sensors makes the size of the target object in the image more suitable, so that a better quality image can be obtained.
  • the target object includes a text area
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image.
  • the hint of the relative position between the image sensors may include:
  • tilt angle of the text line Determine whether the angle of the text line of the text area of the target object relative to one side of the image (hereinafter referred to as the "tilt angle of the text line") is within a specific angle range, if the angle is not located at the specific angle Within the range, the user is prompted to adjust the relative angle between the target object and the image sensor.
  • the specific angle range for example, it may be [-30°, 30°].
  • the specific angle range can be adjusted reasonably, for example, it can be adjusted to: [-20°, 20°], [-10°, 10° ]and many more.
  • the inclination angle of the text line can refer to a certain edge of the text line (for example, the first text line can be selected) relative to the image (for ease of understanding and description, the edge can be defined as a horizontal edge, so The other side of the image adjacent to the side is defined as the angle of the vertical side).
  • “Text line horizontal” may mean that the text line is substantially parallel to the horizontal side of the image (for example, the inclination angle of the text line is less than 30 degrees). If it exceeds the specific angle range, it can be considered that "the text line is inclined”, for example, an angle of about 45 degrees is formed between the text line 806 in FIG. 10 and the horizontal edge of the image. In the case of "the text line is inclined", the user may be prompted to adjust the relative angle between the target object and the image sensor.
  • the inclination angle of a text line can be detected by the following methods: a method based on Hough transform, a method based on cross-correlation, a method based on projection, a method based on Fourier transform, a K-nearest neighbor cluster method, and so on.
  • the relative angle between the target object and the image sensor can be adjusted by adjusting the angle of the target object.
  • the provided relative angle prompt information of the target object may include, but is not limited to, the rotation direction and the rotation angle of the target object. , Can also include only the direction of rotation or the angle of rotation, and can also include other prompt information, for example, you can play a voice such as "Please adjust the angle of the target object".
  • the position and/or posture of the target object in the image is appropriate by judging the inclination angle of the text line of the text area of the target object in the image. For example, judging whether the inclination angle of the text line in the text area of the target object in the image is within a specific angle range can further reflect the image quality, for example, when the inclination angle of the text line in the text area of the target object in the image is not within the specific angle range.
  • the internal time affects the recognition speed and accuracy of the text in the image, and even further affects the services provided to users (such as visually impaired reading services, guidance services, etc.). At this time, it may be necessary to adjust the posture of the target object. Therefore, by prompting the user how to adjust the relative angle between the target object and the image sensor, etc., the position and/or posture of the target object in the image are more suitable, so that a better quality image can be obtained.
  • the image sensor can be made to re-acquire the image and continue to detect until the inclination angle of the text line in the re-acquired image is not less than the specific angle range, Obtain images that meet the quality requirements.
  • Three examples of determining whether to prompt the user to adjust the relative position between the target object and the image sensor based on the geometric relationship between the bounding box and the image are given above, that is, the distance between the edge of the bounding box and the edge of the image (specific distance requirements, For example, it may include a distance requirement based on a distance threshold and a distance requirement based on a distance range), whether the center position of the bounding box is in the central area of the image, and the ratio of the area of the bounding box to the area of the image.
  • an example of judging whether to prompt the user to adjust the relative angle between the target object and the image sensor based on the inclination angle of the text line is also given. Please note that the present disclosure is not limited to the four judgment conditions exemplified above, but other judgment conditions can also be used. Considering the length, the description will not be expanded here.
  • the three judgment conditions given above based on the geometric relationship between the bounding box and the image can be judged separately as described above, or these conditions can be arbitrarily combined for judgment. For example, you can combine them in pairs, for example, first determine whether the center position of the bounding box is in the center area of the image, and then determine whether the area ratio is appropriate; of course, you can also reverse it, first determine whether the area ratio is appropriate, and then determine whether the center position is appropriate Suitable.
  • the three conditions can also be combined with each other. For example, first determine whether the center position of the bounding box is in the center area of the image, and then determine whether the area ratio is appropriate, and finally determine whether the distance between the edges meets the specific distance requirements (for example, within the set distance threshold).
  • judgment of the inclination angle of the text line can also be combined with other conditions.
  • the judgment process of the pairwise combination will be described below in conjunction with examples.
  • at least a part of the target object may be represented by the corresponding bounding box detected in the image.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the distance between the edge of the bounding box and the corresponding edge of the image meets a specific distance requirement
  • step S1042 determines whether the center position of the bounding box is in the center area of the image.
  • step S1043 If it is determined that the center position of the bounding box is not in the center area of the image, then go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • step S1043 if the distance between the edge of the bounding box and the corresponding edge of the image does not meet the specific distance requirement, it can go to step S1043 to provide a method for adjusting the relative position between the target object and the image sensor. prompt.
  • the specific distance requirement may include, for example, the above-mentioned distance requirement based on the distance threshold and/or the distance requirement based on the distance range.
  • the judgment condition of the distance between the bounding box and the edge of the image is combined with the judgment condition of the center position of the bounding box to judge whether the position and/or size of the target object in the image is appropriate.
  • prompts will be automatically provided to the user to help the user quickly adjust the relative position of the target object and the image sensor. Obtain an image that meets the quality requirements (for example, whether the position and/or size of the target object in the image is appropriate), thereby effectively saving processing resources and greatly shortening the processing time.
  • the exemplary embodiment of the present disclosure automatically helps the user to judge the image quality in advance, and further Automatically prompt the user to obtain better quality images to improve the accuracy and speed of target recognition, thereby effectively saving processing resources, greatly shortening the processing time, and helping users to obtain the required services faster (such as reading services) , Guide services, etc.).
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the distance between the edge of the bounding box and the corresponding edge of the image meets a specific distance requirement
  • step S1042 determines whether the ratio of the area of the bounding box to the area of the image is within a specific ratio range ;as well as
  • step S1043 If it is determined that the ratio of the area of the bounding box to the area of the image is not within a specific ratio range, then it may go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • step S1043 if the distance between the edge of the bounding box and the corresponding edge of the image does not meet the specific distance requirement, it can go to step S1043 to provide a method for adjusting the relative position between the target object and the image sensor. prompt.
  • the judgment condition of the distance between the bounding box and the edge of the image is combined with the judgment condition of the area ratio to judge whether the position and size of the target object in the image are appropriate. Similar to the above, when it is determined that the requirements are not met through automatic judgment, as described above, according to some embodiments, prompts will be automatically provided to the user to help the user quickly adjust the relative relationship between the target object and the image sensor. Location, to obtain images that meet quality requirements (such as the appropriate location and size of the target object in the image) faster, thereby effectively saving processing resources and greatly reducing processing time.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the center position of the bounding box is in the center area of the image
  • step S1042 determines whether the ratio of the area of the bounding box to the area of the image is within a specific ratio range
  • step S1043 If it is determined that the ratio of the area of the bounding box to the area of the image is not within the specific ratio range, then go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • step S1043 may go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • the judgment condition of the center position of the bounding box and the judgment condition of the area ratio are combined together to judge whether the position and size of the target object in the image are appropriate. Similar to the above, when it is determined that the requirements are not met through automatic judgment, prompts will be automatically provided to the user to help the user quickly adjust the relative position of the target object and the image sensor, and meet the quality requirements more quickly Image, which effectively saves processing resources and greatly shortens processing time.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the center position of the bounding box is in the center area of the image
  • step S1042 determines whether the distance between the edge of the bounding box and the corresponding edge of the image meets a specific distance requirement
  • step S1043 If it is determined that the distance between the edge of the bounding box and the corresponding edge of the image does not meet the specific distance requirement, go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor .
  • step S1043 may go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • the judgment condition of the center position of the bounding box and the judgment condition of the distance between the bounding box and the edge of the image are combined in order to judge whether the position of the target object in the image is appropriate.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the ratio of the area of the bounding box to the area of the image is within a specific ratio range
  • step S1042 determines whether the distance between the edge of the bounding box and the corresponding edge of the image is within a specific distance Within range
  • step S1043 If it is determined that the distance between the edge of the bounding box and the corresponding edge of the image is not within a specific distance range, then go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor .
  • step S1043 may be transferred to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • the judgment condition of the area ratio and the judgment condition of the center position of the bounding box are combined in order to judge whether the position and size of the target object in the image are appropriate.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the ratio of the area of the bounding box to the area of the image is within a specific ratio range
  • step S1042 If it is determined that the ratio of the area of the bounding box to the area of the image is within a specific ratio range, go to step S1042 to determine whether the center position of the bounding box is in the central area of the image;
  • step S1043 If it is determined that the center position of the bounding box is not in the center area of the image, then go to step S1043 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • step S1043 may be transferred to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • the judgment condition of the area ratio and the judgment condition of the center position of the bounding box are combined in order to judge whether the position of the target object in the image is appropriate.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the distance between the edge of the bounding box and the corresponding edge of the image meets a specific distance requirement
  • step S1042 determines whether the center position of the bounding box is in the center area of the image
  • step S1043 determines whether the ratio of the area of the bounding box to the area of the image is within a certain distance range
  • step S1044 If it is determined that the ratio of the area of the bounding box to the area of the image is not within the specific ratio range, then go to step S1044 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • This embodiment combines the distance of the side, the center position, and the area ratio in order to determine whether the position and size of the target object in the image are appropriate, so as to obtain a target object with a more suitable position and size in the image. .
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 Determine whether the center position of the target object is in the center area of the image
  • step S1042 determines whether the ratio of the area of the bounding box to the area of the image is within a specific distance range
  • step S1043 determines whether the distance between each side of the bounding box and the corresponding side of the image meets the requirement of a certain distance; as well as
  • step S1044 If the distance between each side of the bounding box and the corresponding side of the image does not meet the specific distance requirement, go to step S1044 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • This embodiment combines the area ratio, side distance, and center position in order, and is also used to determine whether the position and size of the target object in the image is appropriate, so as to obtain a target with a more suitable position and size in the image.
  • Object This embodiment combines the area ratio, side distance, and center position in order, and is also used to determine whether the position and size of the target object in the image is appropriate, so as to obtain a target with a more suitable position and size in the image.
  • step S104 in response to the geometric relationship between the at least a part of the target object and the image satisfying a specific condition, it is provided to adjust the target object and the image sensor.
  • the hints of the relative position between can include:
  • Step S1041 determining whether the center position is in the center area of the image
  • step S1042 determines whether the ratio of the area of the bounding box to the area of the image is within a specific ratio range
  • step S1043 determines whether the distance between each side of the bounding box and the corresponding side of the image meets the specific distance requirement ;as well as
  • step S1044 If it is determined that the distance between each side of the bounding box and the corresponding side of the image does not meet the specific distance requirement, then it may go to step S1044 to provide a prompt for adjusting the relative position between the target object and the image sensor.
  • This embodiment combines the center position, area ratio, and edge distance in order, and is also used to determine whether the position and size of the target object in the image is appropriate, so as to obtain a target with a more suitable position and size in the image.
  • This combination method can more quickly obtain the target object with suitable position and size in the image.
  • the response to the geometric relationship between the at least a part of the target object and the image satisfies a specific condition, and a prompt for adjusting the relative position between the target object and the image sensor may also be provided.
  • step S1041' it is determined whether the angle of the text line of the text area of the target object relative to an edge of the image is within a specific angle range;
  • the user may be prompted to adjust the relative relationship between the target object and the image sensor in step S1042' angle.
  • the specific angle range may be, for example, [-30°, 30°] and the relative angle may include, for example, two parameters of rotation angle and rotation direction.
  • FIG. 11D may also include the steps shown in FIGS. 10A to 10F and 11A to 11C.
  • steps S1041' and S1042' can be combined with the steps in FIGS. 10A to 10F and FIGS. 11A to 11C, respectively. Due to space limitations, the narrative will not be expanded here.
  • Step S104 has been described in detail above, and other steps will be further described below.
  • step S103 provides a prompt to adjust the target object to the field of view of the image sensor, as shown in FIG. 1, you can go to step S101 to reacquire the image re-acquired by the image sensor and perform step S101.
  • step S102 continues to detect whether at least a part of the target object exists in the reacquired image until at least a part of the target object is detected in the reacquired image.
  • step S104 provides a prompt for adjusting the relative position between the target object and the image sensor, as shown in FIG. 1, you can also go to step S101 to reacquire the image reacquired by the image sensor and based on the reacquisition Continue to perform steps S102, S104, etc., until at least a part of the target object is detected in the reacquired image, and it is determined that the geometric relationship between at least a part of the detected target object and the reacquired image no longer meets the requirements
  • the image can be re-acquired to detect whether there is a target object and the existing target in the re-acquired image Whether the size and/or location of the object are appropriate, etc.
  • the number and/or time of re-acquisition and detection and prompting can be controlled.
  • the user can actively (artificially) start the stop detection instruction; or the processor and other devices can start the stop detection instruction by itself.
  • it can stop the continued image acquisition based on the number of detections, the time when the bounding box is not continuously detected, etc. , Continue to detect, and stop the prompts issued because the bounding box or the complete bounding box is not detected, or because the relative position of the target object and the image sensor is continuously inappropriate, etc., so as to avoid too concentrated and frequent prompts, and provide users with Services that meet demand.
  • the image sensor can be made to re-acquire the target object after at least a specific time (such as 2s), so that there is enough time to move the target object and/or the device where the image sensor is located according to the prompts to prevent When the relative position of the target object and the image sensor is prepared, the target object is collected too frequently.
  • a specific time such as 2s
  • the realization of re-acquisition of the target object after a specific time interval between the image sensor can be achieved, for example, by using interval collection, that is, collecting at a fixed time interval, such as a shutter every 2 seconds. This interval can be set in some image sensors. Collection method.
  • step S110 the target object to be acquired is confirmed by detecting the guide object, so that the target object is confirmed before the acquisition to prevent the image
  • the target object is confirmed before the acquisition to prevent the image
  • the pointer may be a user's finger, for example.
  • computer vision algorithms can be used to detect the finger and determine the position of the finger.
  • an algorithm from OpenCV can be used, such as the CascadeClassifier function (https://docs.opencv.org/3.2.0/d1/de5/classcv_1_1CascadeClassifier. html) to determine the position of the finger from the image to confirm the target object.
  • the guide is not limited to a finger, for example, it may also be a pen.
  • the user may also be prompted to confirm the target object to be collected in the form of text or voice prompts.
  • the target detection method of the present disclosure may further include: step S106, recognizing the target object in the image and outputting the recognized target object. For example, in response to detecting at least a part of the target object in the image, and determining that the geometric relationship between at least a part of the detected target object and the image does not satisfy the need, it is provided to adjust the relative relationship between the target object and the image sensor.
  • the specific condition of the position prompt, or the condition that does not need to provide a prompt for adjusting the relative position between the target object and the image sensor that is, the size and/or position of the target object in the image is more appropriate, and no adjustment is required.
  • the prompt of the relative position between the target object and the image sensor can recognize the target object in the image and output the recognized target object.
  • the target object may be a movable target object, such as an ID card, business card, passport, driving license, reading material (media reading material or electronic reading material), Tablet computers, mobile phones, etc.
  • a movable target object such as an ID card, business card, passport, driving license, reading material (media reading material or electronic reading material), Tablet computers, mobile phones, etc.
  • recognizing the target object in the image may include recognizing text in the target object, and accordingly, outputting the recognized target object may include outputting text in the recognized target object.
  • the output format includes audio output and/or text output.
  • the exemplary embodiments of the present disclosure automatically help the user to judge the image quality in advance, and further automatically prompt the user to obtain Better quality images to improve the accuracy and speed of target recognition, thereby effectively saving processing resources, greatly shortening the processing time, and helping users get the services they need faster (such as reading services, guidance services, etc.) ).
  • the image sensor may first collect a preview image to detect the target object based on the preview image, and after detecting the bounding box of the target object or the complete bounding box of the target object, the image for recognition (ie Photographed image), wherein the resolution of the preview image is smaller than the resolution of the image used for recognition.
  • the image for recognition ie Photographed image
  • a prompt to maintain the target object may be provided, so that the user maintains the position of the target object, or if the user Need to hold or support the target object, so that the user can prepare in advance, hold or hold it more firmly, so as to take a clear image.
  • the above-mentioned prompt may include one of a sound prompt, a vibration prompt, a text prompt, an image prompt, and a video prompt, or any combination thereof. That is to say, the present disclosure is not limited to a certain prompt method, but can be combined in a variety of ways to prompt.
  • Vibration notification is very convenient for hearing impaired people who cannot hear normal sounds.
  • the present disclosure proposes to first detect whether there is at least a part of the target object in the image collected by the image sensor, and then provide if at least a part of the target object is not detected
  • the relative position of the sensor is prompted to more conveniently adjust the position of the target object and/or the image sensor, so that the image sensor can quickly collect the target object with a suitable position and/or size.
  • the exemplary embodiments of the present disclosure automatically help the user to judge the image quality in advance, and further automatically prompt Users can obtain better quality images to improve the accuracy and speed of target recognition, thereby effectively saving processing resources, greatly shortening processing time, and helping users to obtain required services faster (such as reading services, guidance Services etc.).
  • an electronic circuit which may include: a circuit configured to perform the steps of the above-mentioned method.
  • a visually impaired auxiliary device including: an image sensor configured to capture an image; the above-mentioned electronic circuit, for example, the electronic circuit is configured to detect whether there is a target object in the image At least a part of and in response to at least a part of the target object not being detected in the image, provide a prompt for adjusting the target object to within the field of view of the image sensor.
  • the electronic circuit may be configured to provide a prompt to move the target object so that the target object enters the field of view of the image sensor in response to the image sensor being a fixed-position device. And in response to the image sensor being a position-adjustable device, providing a prompt to move the target object and/or the image sensor so that the target object enters the field of view of the image sensor.
  • the visually impaired assistive device may include one or more of the following devices: a visually impaired assistive device (e.g., reading assistive device, etc.) that can be connected to a wearable device (e.g., glasses, helmet, etc.), Hand-held assistive devices for the visually impaired and desktop assistive devices for the visually impaired, etc.
  • a visually impaired assistive device e.g., reading assistive device, etc.
  • a wearable device e.g., glasses, helmet, etc.
  • Hand-held assistive devices for the visually impaired e.g., glasses, helmet, etc.
  • the visually impaired auxiliary device may be installed on a corresponding stand, such as a desktop stand, as a desktop-type visually impaired auxiliary device.
  • the visually impaired auxiliary device may further include a corresponding bracket.
  • FIGS. 14A and 14B respectively show a structural block diagram of a stand for a visually impaired assist device in an unfolded state and a folded state, such as a desktop stand, according to an exemplary embodiment of the present disclosure.
  • the bracket may include a connecting piece 141, a telescopic stand arm 142, a base rotating joint 143 and a base 144 with a counterweight.
  • the connecting member 141 is configured to mount (for example, by suction mounting) the image sensor and the electronic circuit.
  • the connecting member 141 may include a magnet embedded and formed on the surface of the connecting member 141 (for example, when the bracket is stabilized on the desktop in an unfolded state, the connecting member 141 faces the desktop The surface) is used to attach the image sensor and the electronic circuit to the connector 141 by suction.
  • the telescopic stand arm 142 may include a plurality of telescopic sections connected to each other, for example, four telescopic sections as shown in FIG. 14A, of course, other numbers of telescopic sections may also be included according to actual requirements. The length of at least one telescopic section of the plurality of telescopic sections is adjustable.
  • the position of the bottom telescopic joint connected to the base rotating joint 143 among the plurality of telescopic joints can be adjusted relative to the weight base 144 (for example, it can be rotated within a set angle range relative to the weight base 144,
  • the set angle range may be, for example, 0-90° including the end value, etc.).
  • Other expansion joints can be sleeved on the previous expansion joint connected to it, or can be adjusted relative to the previous expansion joint connected to it (for example, can be rotated relative to the previous expansion joint within a set angle range, the set angle
  • the range may be 0 to 90° including the end value, etc.).
  • the counterweight base 144 is configured to support other components of the bracket, such as the connecting piece 141 and the telescopic stand arm 142, so that the bracket can be stably on the desktop when it is in the unfolded state (or working state).
  • the base with counterweight 144 can stabilize the bracket on the desktop.
  • the cross-section of the weighted base 144 may be an approximately rectangular shape with four arc-shaped ends, wherein the long side of the rectangle may have a size of 150-180 mm, as shown in FIG.
  • the short side of the rectangle may have a size of 60-90mm, such as 80mm shown in the figure.
  • the cross-section of the counterweight base 144 can also be set to other shapes, such as a circle, an ellipse, and so on.
  • the base rotation joint 143 may be located on the edge of one side of the base 144 with counterweight.
  • the bracket when the bracket is in the unfolded state, it can be more Good supporting effect, on the other hand, when the bracket is in the folded state, the volume can be further reduced, and the user experience can be enhanced.
  • FIG. 14A and FIG. 14B A possible structure of a stand that can be used for the visually impaired auxiliary device, such as a desktop stand, is described above schematically according to Figs. 14A and 14B. It should be noted that the structure shown in FIG. 14A and FIG. 14B is only an example. According to a specific implementation, the stent of the present disclosure may only include one or more of the components shown in FIG. 14A or FIG. 14B. One or more components can be included.
  • the visually impaired assist device may further include a circuit configured to perform text detection and recognition on the text contained in the image to obtain text data; and a circuit configured to convert the text data into sound data. Circuit; and a circuit configured to output the sound data and/or the text data, and so on.
  • an electronic device including: a processor; and a memory storing a program, the program including instructions that, when executed by the processor, cause the processor to execute the foregoing Methods.
  • a non-transitory computer-readable storage medium storing a program, the program including instructions that, when executed by a processor of an electronic device, cause the electronic device to execute the foregoing Methods.
  • FIG. 15 is a block diagram showing an example of an electronic device according to an exemplary embodiment of the present disclosure. It should be noted that the structure shown in FIG. 15 is only an example, and according to specific implementation manners, the electronic device of the present disclosure may only include one or more of the components shown in FIG. 15.
  • the electronic device 2000 may be, for example, a general-purpose computer (such as a laptop computer, a tablet computer, etc.), a mobile phone, or a personal digital assistant. According to some embodiments, the electronic device 2000 may be a visually impaired assist device.
  • the electronic device 2000 may be configured to capture or capture images, process the captured (for example, preview) images, and provide prompts in response to the processing.
  • the electronic device 2000 may be configured to collect an image, perform text detection and recognition on the image to obtain text data, convert the text data into sound data, and output sound data for the user to listen to, and/or output text data for the user Watch.
  • the electronic device 2000 may be configured to include a spectacle frame or be configured to be detachably mountable to a spectacle frame (for example, a frame of a spectacle frame, a connector connecting two spectacle frames, temples, or any other part). ), so that an image that approximately includes the user's field of view can be collected or photographed.
  • a spectacle frame for example, a frame of a spectacle frame, a connector connecting two spectacle frames, temples, or any other part.
  • the electronic device 2000 may also be installed on other wearable devices, or integrated with other wearable devices.
  • the wearable device may be, for example, a head-mounted device (such as a helmet or a hat, etc.), a device that can be worn on the ear, and the like.
  • the electronic device can be implemented as an accessory that can be attached to a wearable device, for example, can be implemented as an accessory that can be attached to a helmet or a hat, or the like.
  • the electronic device 2000 may also have other forms.
  • the electronic device 2000 may be a mobile phone, a general-purpose computing device (such as a laptop computer, a tablet computer, etc.), a personal digital assistant, and so on.
  • the electronic device 2000 may also have a base so that it can be placed on a desktop.
  • the electronic device 2000 can be used as a visually impaired auxiliary device to assist reading.
  • the electronic device 2000 is sometimes called an "electronic reader” or a “reading auxiliary device”.
  • users who are unable to read autonomously can use postures similar to reading postures to achieve "reading” of conventional reading materials (such as books, magazines, etc.).
  • the electronic device 2000 may acquire the captured image, detect the image, and determine whether there is a bounding box surrounding the reading in the image to confirm whether the reading is collected into the Image, and if it is determined that there is a bounding box, but the size or position of the collected reading is not suitable, the user can be prompted to move the reading based on the position of the bounding box in the image to obtain a reading with a suitable size and position. And perform text detection and text recognition (for example, using the OCR method of optical text recognition) on the reading material of the appropriate size and position to obtain the text data, and can convert the text data into sound data, and output it through speakers or headphones. The device outputs the sound data for the user to listen to.
  • text detection and text recognition for example, using the OCR method of optical text recognition
  • the electronic device 2000 can help the electronic device 2000 to obtain target objects with suitable positions and sizes faster and more accurately, so as to be able to Faster and more accurate recognition and output to users can greatly improve response speed, recognition efficiency and recognition accuracy.
  • the electronic device 2000 may include an image sensor 2004 for collecting, photographing, and acquiring images.
  • the image sensor 2004 may collect and/or shoot static images, and may also collect and/or shoot dynamic images, and may include but is not limited to cameras, cameras, video cameras, etc., and is configured to obtain an initial image including a target object.
  • the electronic device 2000 may also include an electronic circuit 2100 that includes a circuit configured to perform the steps of the method as previously described.
  • the electronic device 2100 may further include a text recognition circuit 2005, which is configured to perform text detection and recognition (for example, OCR processing) on the text in the image, so as to obtain text data.
  • the character recognition circuit 2005 can be implemented by a dedicated chip, for example.
  • the electronic device 2000 may further include a sound conversion circuit 2006 configured to convert the text data into sound data.
  • the sound conversion circuit 2006 may be realized by a dedicated chip, for example.
  • the electronic device 2000 may further include a sound output circuit 2007 configured to output the sound data.
  • the sound output circuit 2007 may include, but is not limited to, earphones, speakers, or vibrators, etc., and their corresponding driving circuits.
  • the electronic device 2000 may further include an image processing circuit 2008, and the image processing circuit 2008 may include a circuit configured to perform various image processing on an image.
  • the image processing circuit 2008 may, for example, include, but is not limited to, one or more of the following: a circuit configured to denoise an image, a circuit configured to defuzzify an image, a circuit configured to perform geometric correction on an image A circuit, a circuit configured to perform feature extraction on an image, a circuit configured to perform target detection and recognition of a target object in an image, a circuit configured to perform text detection on a text contained in an image, a circuit configured to perform a text detection from an image A circuit configured to extract text lines in an image, a circuit configured to extract text coordinates from an image, a circuit configured to extract a bounding box from an image, a circuit configured to extract a text box from an image, and a circuit configured to perform layout based on an image Analyze (for example, paragraph division) of the circuit, and so on.
  • the electronic circuit 2100 may further include a word processing circuit 2009, which may be configured to be based on the extracted text-related information (such as text data, text boxes, paragraph coordinates, text line coordinates, (Text coordinates, etc.) to perform various processing to obtain processing results such as paragraph sorting, text semantic analysis, and layout analysis results.
  • a word processing circuit 2009 may be configured to be based on the extracted text-related information (such as text data, text boxes, paragraph coordinates, text line coordinates, (Text coordinates, etc.) to perform various processing to obtain processing results such as paragraph sorting, text semantic analysis, and layout analysis results.
  • One or more of the above-mentioned various circuits may use customized hardware, and/or may It is implemented by hardware, software, firmware, middleware, microcode, hardware description language, or any combination thereof.
  • one or more of the above-mentioned various circuits can be implemented in assembly language by using logic and algorithms according to the present disclosure.
  • a hardware programming language such as VERILOG, VHDL, C++
  • programming hardware for example, a programmable logic circuit including a field programmable gate array (FPGA) and/or a programmable logic array (PLA)).
  • the electronic device 2000 may further include a communication circuit 2010, which may be any type of device or system that enables communication with external devices and/or with a network, and may include, but is not limited to, a modem, a network card , Infrared communication devices, wireless communication devices and/or chipsets, such as Bluetooth devices, 1302.11 devices, WiFi devices, WiMax devices, cellular communication devices and/or the like.
  • a communication circuit 2010, may be any type of device or system that enables communication with external devices and/or with a network, and may include, but is not limited to, a modem, a network card , Infrared communication devices, wireless communication devices and/or chipsets, such as Bluetooth devices, 1302.11 devices, WiFi devices, WiMax devices, cellular communication devices and/or the like.
  • the electronic device 2000 may further include an input device 2011.
  • the input device 2011 may be any type of device that can input information to the electronic device 2000, and may include, but is not limited to, various sensors, a mouse, a keyboard, and a touch screen. , Buttons, joystick, microphone and/or remote control, etc.
  • the electronic device 2000 may further include an output device 2012.
  • the output device 2012 may be any type of device capable of presenting information, and may include, but is not limited to, a display, a visual output terminal, a vibrator, and/or a printer, etc. .
  • the electronic device 2000 is used as an assistive device for the visually impaired according to some embodiments, the vision-based output device can facilitate the user's family or maintenance staff to obtain output information from the electronic device 2000.
  • the electronic device 2000 may further include a processor 2001.
  • the processor 2001 may be any type of processor, and may include, but is not limited to, one or more general-purpose processors and/or one or more special-purpose processors (for example, special processing chips).
  • the processor 2001 may be, but is not limited to, a central processing unit CPU or a microprocessor MPU, for example.
  • the electronic device 2000 may also include a working memory 2002, which may store programs (including instructions) and/or data (such as images, text, sounds, and other intermediate data, etc.) useful for the work of the processor 2001.
  • Memory and may include, but is not limited to, random access memory and/or read-only memory devices.
  • the electronic device 2000 may also include a storage device 2003.
  • the storage device 2003 may include any non-transitory storage device.
  • the non-transitory storage device may be any storage device that is non-transitory and can realize data storage, and may include but is not limited to Disk drives, optical storage devices, solid-state memory, floppy disks, flexible disks, hard disks, tapes or any other magnetic media, optical disks or any other optical media, ROM (read only memory), RAM (random access memory), cache memory, and /Or any other memory chip or cartridge, and/or any other medium from which the computer can read data, instructions and/or code.
  • the working memory 2002 and the storage device 2003 may be collectively referred to as "memory", and in some cases may be used with each other.
  • the processor 2001 can provide information to the image sensor 2004, the character recognition circuit 2005, the sound conversion circuit 2006, the sound output circuit 2007, the image processing circuit 2008, the word processing circuit 2009, the communication circuit 2010, the electronic circuit 2100, and the electronic device 2000. At least one of the other various devices and circuits included is controlled and scheduled. According to some embodiments, at least some of the various components described in FIG. 15 may be connected and/or communicated with each other through a line 2013.
  • Software elements may be located in the working memory 2002, including but not limited to an operating system 2002a, one or more application programs 2002b, drivers, and/or other data and codes.
  • instructions for performing the aforementioned control and scheduling may be included in the operating system 2002a or one or more application programs 2002b.
  • the instructions for executing the method steps described in the present disclosure may be included in one or more application programs 2002b, and the various modules of the above-mentioned electronic device 2000 may be read and executed by the processor 2001.
  • the instructions of the application program 2002b are implemented.
  • the electronic device 2000 may include a processor 2001 and a memory (for example, a working memory 2002 and/or a storage device 2003) storing a program.
  • the program includes instructions that, when executed by the processor 2001, cause the processing
  • the device 2001 executes methods as described in various embodiments of the present disclosure.
  • part or all of the operations performed by at least one of the character recognition circuit 2005, the sound conversion circuit 2006, the image processing circuit 2008, the word processing circuit 2009, and the electronic circuit 2100 may be read and executed by the processor 2001
  • One or more instructions of the application program 2002 are implemented.
  • the executable code or source code of the instructions of the software element (program) can be stored in a non-transitory computer-readable storage medium (such as the storage device 2003), and can be stored in the working memory 2001 (which may be Compile and/or install). Therefore, the present disclosure provides a computer-readable storage medium storing a program.
  • the program includes instructions that, when executed by a processor of an electronic device (for example, a visually impaired assist device), cause the electronic device to execute the steps of the present disclosure.
  • the executable code or source code of the instructions of the software element (program) can also be downloaded from a remote location.
  • the processor 2001 in the electronic device 2000 may be distributed on a network.
  • one processor may be used to perform some processing, while at the same time another processor remote from the one processor may perform other processing.
  • Other modules of the electronic device 2001 can also be similarly distributed. In this way, the electronic device 2001 can be interpreted as a distributed computing system that performs processing in multiple locations.

Abstract

提供一种目标检测方法、电路、视障辅助设备、电子设备和介质。所述目标检测方法包括：获取图像传感器采集的图像；检测所述图像中是否存在目标对象的至少一部分；响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。

Description

目标检测方法、电路、视障辅助设备、电子设备和介质 技术领域
本公开涉及图像处理领域，特别涉及一种目标检测方法、电子电路、视障辅助设备、电子设备和计算机可读存储介质。
背景技术
近年来，目标检测技术在各领域得到了广泛的应用，一直是业界关注的焦点之一。
在此部分中描述的方法不一定是之前已经设想到或采用的方法。除非另有指明，否则不应假定此部分中描述的任何方法仅因其包括在此部分中就被认为是现有技术。类似地，除非另有指明，否则此部分中提及的问题不应认为在任何现有技术中已被公认。
发明内容
根据本公开的一方面，提供一种目标检测方法，包括：获取图像传感器采集的图像；检测所述图像中是否存在目标对象的至少一部分；以及响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。
根据本公开的另一方面，提供一种电子电路，包括：被配置为执行上述的方法的步骤的电路。
根据本公开的另一方面，还提供一种视障辅助设备，包括：图像传感器，被配置为采集图像；以及电子电路，被配置为检测所述图像中是否存在目标对象的至少一部分，其中，所述电子电路还被配置为响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。
根据本公开的另一方面，还提供一种电子设备，包括：处理器；以及存储程序的存储器，所述程序包括指令，所述指令在由所述处理器执行时使所述处理器执行上述的方法。
根据本公开的另一方面，还提供一种存储程序的非暂态计算机可读存储介质，所述程序包括指令，所述指令在由电子设备的处理器执行时，致使所述电子设备执行上述的方法。
附图说明
附图示例性地示出了实施例并且构成说明书的一部分，与说明书的文字描述一起用于讲解实施例的示例性实施方式。所示出的实施例仅出于例示的目的，并不限制权利要求的范围。在所有附图中，相同的附图标记指代类似但不一定相同的要素。
图1是示出根据本公开的一个示例性实施例的目标检测方法的流程图。
图2A示出了在图像中检测到的包含目标对象的一部分的边界框的示意图，图2B示出了在图像中检测到的包含整个目标对象的边界框的示意图。
图3示出了一个检测图像中是否存在目标对象的至少一部分的示例性过程的示意图。
图4示出了一个示例性的用于通过距离阈值来判断是否是完整边界框的图示。
图5A、图5B、图5C、图5D、图5E、图5F示出了边界框的边与图像的边出现重合的示例情形。
图6给出了一个基于边界框的边与图像的相应边的距离范围进行判断的示例。
图7A、图7B、图7C示出了边界框的边与图像的相应边的距离不在特定距离范围内的示例情形。
图8A、图8B给出了边界框的中心位置与图像的中心区域之间的位置关系的示例情形。
图9A、图9B、图9C给出边界框的面积与图像的面积之间的关系的示例情形。
图10给出了文本行倾斜的示例情形。
图11A、图11B、图11C、图11D、图11E、图11F示出了两两组合的判断方式的示例。
图12A、图12B、图12C示出了组合三种判断条件来进行判断的示例。
图12D示出了基于文本行的角度进行判断的示例。
图13是示出根据本公开的另一个示例性实施例的目标检测方法的流程图。
图14A和图14B分别示出了根据本公开的示例性实施例的用于视障辅助设备的、处于展开状态和折叠状态的支架例如桌面式支架的结构框图。
图15是示出根据本公开的示例性实施例的电子设备的结构框图。
具体实施方式
在本公开中，除非另有说明，否则使用术语“第一”、“第二”等来描述各种要素不意图限定这些要素的位置关系、时序关系或重要性关系，这种术语只是用于将一个元件与 另一元件区分开。在一些示例中，第一要素和第二要素可以指向该要素的同一实例，而在某些情况下，基于上下文的描述，它们也可以指代不同实例。
在本公开中对各种所述示例的描述中所使用的术语只是为了描述特定示例的目的，而并非旨在进行限制。除非上下文另外明确地表明，如果不特意限定要素的数量，则该要素可以是一个也可以是多个。此外，本公开中所使用的术语“和/或”涵盖所列出的项目中的任何一个以及全部可能的组合方式。
对于不提供预览图(例如不带显示屏的可穿戴设备)或者无法看到预览图(例如视障用户)的情形，会存在图像传感器未对准目标对象的情况，这种未对准的情况导致目标对象的识别存在问题。例如，在目标对象中包含文字的情况下，这种未对准的情况使得文字识别出现问题。
本公开提供了一种目标检测方法。图1是示出根据本公开的示例性实施例的目标检测方法的流程图。
如图1所示，根据本公开的示例性实施例的目标检测方法包括：步骤S101，获取图像传感器采集的图像；步骤S102，检测所述图像中是否存在目标对象的至少一部分；步骤S103，响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。
在步骤S101，获取图像传感器采集的图像，以便后续可检测图像中是否存在目标对象的至少一部分。
根据一些实施例，用于采集图像的图像传感器能够进行静态或动态的图像采集，其可以是独立装置(例如照相机、视频摄像机、摄像头等)，也可以包括在各类电子设备(例如移动电话、计算机、个人数字助理、视障辅助设备、平板计算机、阅读辅助设备、可穿戴设备等)中。
根据一些实施例，在本步骤中，图像传感器采集的图像可以是预览图像，而不必须是拍照图像。
根据一些实施例，图像传感器可以设置于例如用户的可穿戴设备、辅助阅读眼镜、手持式设备等设备上，从而获取的所述图像可以是由例如在可穿戴设备或辅助阅读眼镜上设置的图像传感器所采集的目标对象的图像。
根据一些实施例，目标对象可以是放置在某处的物体，或者可以是握持在用户手中或者由用户扶持的物体等。
根据一些实施例，目标对象可以具有规则的形状，也可以具有不规则的形状。
根据一些实施例，目标对象可以是不固定的，例如可以由用户移动，也可以通过用于移动目标对象的移动装置来移动，比如能够移动的目标对象可以为：身份证、名片、护照、驾照、读物、平板计算机、移动电话等；甚至，目标对象自身可以移动，例如车辆等。根据一些实施例，目标对象也可以是不能移动的，例如安装在墙上的电视、固定的公告牌等。
根据一些实施例，所述目标对象中可包含例如各种形式的文字、数字、字符、符号、图等内容。
根据一些实施例，所获取的图像可以直接是由图像传感器采集的图像，也可以是在图像传感器采集的图像基础上经过了某种或一些预先处理的图像，所述预先处理例如可以包括去躁、对比度增强、分辨率处理等等。
根据一些实施例，可以实时地从图像传感器获取图像，也可以在图像传感器采集图像之后的一段时间获取图像。
根据一些实施例，获取的图像可以是经过预先筛选的图像，例如经过多次采集，选取其中较为清楚的图像。
在步骤S102，在获取图像之后，检测所述图像中是否存在目标对象的至少一部分。
根据一些实施例，可通过对所述图像进行边界框检测来确定所述图像中是否存在目标对象的至少一部分。例如，响应于在所述图像中未检测到边界框，可确定在所述图像中未检测到目标对象的至少一部分，响应于在所述图像中检测到边界框，可确定在所述图像中能够检测到目标对象的至少一部分。
根据一些实施例，可以使用目标检测算法(例如R-CNN、Fast R-CNN、RFCN等)来检测目标对象的边界框。
目标检测算法在检测到图像中存在目标对象的至少一部分之后，可以输出用于表示相应边界框的位置的参数，这样的参数可以包括该边界框的中心坐标、宽度和高度等的信息。由此可以根据得到的边界框的中心坐标、宽度和高度等的信息，获取边界框的各顶点的坐标，以得到相应的边界框。请注意，用于表示边界框的位置的参数还可以包括其它能够表示位置信息的参数组合，而不仅限于这里举出的示例。实际上，可以根据目标检测的具体需求来选择要使用哪些用于表示位置的参数。
根据一些实施例，目标对象的边界框检测还可以采用其它的用于根据目标对象的位置来确定边界框的检测方式。
如图1所示，响应于在所述图像中未检测到目标对象的至少一部分，则可以执行步骤S103，提供调整所述目标对象至所述图像传感器视场范围内的提示。
由此可见，在本公开中，可首先判断是否存在检测不到目标对象的至少一部分的情况(例如，检测不到边界框的情况)，以自动地判断图像质量，并且如果确定存在这种情况，还将自动地给出适合用户的提示(例如，调整所述目标对象至所述图像传感器视场范围内的提示)，以有助于用户快速地调整目标对象与图像传感器的相对位置，更快地获得满足质量要求(例如目标对象在图像中的位置和/或大小等是否合适)的图像，从而有效地节约了处理资源，大大地缩短了处理时间。
根据一些实施方式，本公开的目标检测方法和相关电路及设备等可以用于视障辅助和可穿戴设备，在视力障碍的用户难以用视觉判断图像质量或者用户佩戴的可穿戴设备不提供显示屏的情况下，通过自动帮助用户预先判断图像质量，并且还能进一步自动提示用户，从而获得更好质量的图像，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
根据一些实施例，在步骤S103中，提供调整所述目标对象至所述图像传感器视场范围内的提示，可包括：响应于所述图像传感器为位置固定的设备，提供移动所述目标对象以使所述目标对象进入所述图像传感器视场范围内的提示；以及响应于所述图像传感器为位置可调的设备，提供移动所述目标对象和/或所述图像传感器以使所述目标对象进入所述图像传感器视场范围内的提示。
示例的，假设所述图像传感器被安装于位置相对固定的桌面式视障辅助设备等设备中，即，该图像传感器的位置相对固定，则可提供移动所述目标对象以使所述目标对象进入所述图像传感器视场范围内的提示，以便于用户操作，提高用户的使用体验。假设所述图像传感器被安装于位置较为灵活可调的手持式视障辅助设备、或者可穿戴式视障辅助设备等设备中，即，该图像传感器的位置相对灵活可调，则不仅可提供移动所述目标对象以使所述目标对象进入所述图像传感器视场范围内的提示，还可提供移动所述图像传感器以使所述目标对象进入所述图像传感器视场范围内的提示，以增加用户操作的灵活性，提高用户的使用体验。
根据一些实施例，在步骤S103中，除了可提供调整所述目标对象至所述图像传感器视场范围内的提示之外，还可同时提供检测不到目标对象的至少一部分的提示。根据前述相关描述，可通过对所述图像进行边界框检测来确定所述图像中是否存在目标对象的至少一部分，因而，所述检测不到目标对象的至少一部分的提示可以包括“检测不到边 界框”，“检测不到包含目标对象的至少一部分的边界框”，或者“确定在图像中不存在包含目标对象的至少一部分的边界框”的提示，等等。
根据一些实施例，提供移动所述目标对象的提示可以包括提供移动所述目标对象使其远离和/或偏移的提示，提供移动所述图像传感器的提示可以包括提供移动所述图像传感器(例如，移动所述图像传感器本身，或者移动所述图像传感器所在设备以移动所述图像传感器)使其远离和/或偏移的提示。由于可通过确定检测不到边界框来确定在所述图像中未检测到目标对象的至少一部分，而检测不到边界框可能是因为目标对象太近、或者相对于图像传感器的取景框位置(或视场范围)有些偏，所以可以通过移动目标对象(和/或图像传感器)使其远离和/或偏移来使得所述目标对象能够尽可能进入所述图像传感器的视场范围内。另外，值的注意的是，所述图像传感器的视场范围接近但不同于用户的视场范围，相比较而言，调整所述目标对象至所述图像传感器的视场范围内相对于调整所述目标对象至相应用户的视场范围内会使得目标检测的结果更为精确。
如上所述，对于检测不到边界框从而确定在所述图像中未检测到目标对象的至少一部分的情形，可以调整目标对象与图像传感器的远近以更多进入取景框的视场，或者调整目标对象和/或图像传感器的方向(上下左右的方向)来更多进入取景框的视场，当然，在远近和方向上都调整也是可以的。
根据一些实施例，对于检测不到边界框从而确定在所述图像中未检测到目标对象的至少一部分的情况，可以优选地调整目标对象的远近。
根据一些实施例，如图1所示，所述方法还可包括步骤S104：响应于在所述图像中检测到所述目标对象的至少一部分，确定所述目标对象的所述至少一部分与所述图像的几何关系是否满足特定条件；以及响应于所述目标对象的所述至少一部分与所述图像的几何关系满足所述特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置、以使得所述目标对象位于所述图像传感器视场范围内的适当位置的提示。
如前所述，响应于在所述图像中检测到边界框，可确定在所述图像中能够检测到目标对象的至少一部分。根据一些实施例，检测到的边界框可以为包含目标对象的一部分的边界框，或者包含整个目标对象的边界框。
示例的，图2A示出了在图像801中检测到的、其中包含目标对象802的一部分的边界框803(不完整边界框)，图2B示出了在图像801’中检测到的、其中包含整个目标对象802的边界框803’(完整边界框)。
在图像中检测到的目标对象的边界框至少包围被采集到图像中的目标对象(图像中可能包含整个目标对象，也有可能只包含目标对象的一部分)，也就是说，目标对象的边界框至少包围由目标对象在图像中的实际轮廓构成的真实边框。
虽然图2A中示出的边界框803包围的区域不仅包括目标对象802的一部分，还包括采集到的目标对象802之外的区域，但是，边界框803也可以与采集到的目标对象802的那一部分的真实轮廓相吻合(大致吻合)，此时的边界框为包围目标对象802的那一部分的真实轮廓的最小边界框。类似地，虽然图2B中示出的完整边界框803’包围的区域不仅包括整个目标对象802，还包括目标对象802之外的区域，但是，完整边界框803’也可以与目标对象802的整个真实轮廓相吻合(大致吻合)，此时的完整边界框为包围目标对象802的整个真实轮廓的最小边界框。
换言之，本文所述的“边界框”可以指在该边界框中包括目标对象的至少一部分，即，可以是全部，也可以是一部分；而“完整边界框”可以指在该边界框中要包含整个目标对象。如果没有检测到边界框，则属于“检测不到包含目标对象的至少一部分的边界框”或者“确定在图像中不存在包含目标对象的至少一部分的边界框”。如果没有将整个目标对象而是将目标对象的部分采集进图像，导致在图像中检测不到目标对象的完整边界框，则属于检测到边界框，但“检测不到包含整个目标对象的完整边界框”或者“确定在图像中不存在包含整个目标对象的完整边界框”的情况。
根据一些实施例，响应于在所述图像中检测到所述目标对象的至少一部分之后，还可进一步确定检测到的所述目标对象的至少一部分是否包含了整个目标对象，即，是否能够检测到完整目标对象。例如，在检测到目标对象的边界框(即确定图像中存在目标对象的至少一部分)之后，再检测该边界框是否属于完整边界框，即该边界框是否包含了整个目标对象，以确定是否能够检测到完整目标对象。
示例的，如图3所示，步骤S102中的所述“检测图像中是否存在目标对象的至少一部分”例如可以包括步骤S1021、S1022。
在步骤S1021，检测所述图像中是否存在包含目标对象的至少一部分的边界框。
如前所述，可以使用目标检测算法(例如R-CNN、Fast R-CNN、RFCN等)来检测目标对象的边界框。目标检测算法在检测到图像上的目标对象之后，可以输出用于表示边界框的位置的参数，这样的参数可以包括该边界框的中心坐标、宽度和高度的信息。由此可以根据得到的边界框的中心坐标、宽度和高度的信息，获取边界框的各顶点的坐标。
根据一些实施例，检测到的目标对象的边界框可以是一个矩形框，如图2A与图2B所示。此时，该矩形框例如可以通过四个顶点位置来确定。
根据一些实施例，检测到的目标对象的边界框还可以是矩形以外的形状，例如圆形、多边形、椭圆形等其他规则的形状，或者不规则的各种形状。本公开对于边界框的形状不加限制。
根据一些实施例，如果在步骤S1021中，在所述图像中未检测到所述边界框(即确定在所述图像中未检测到所述目标对象的至少一部分)，则转到步骤S103，提供调整所述目标对象至所述图像传感器视场范围内的提示。另外，同时也可提供检测不到边界框的提示。所述检测不到边界框的提示属于检测不到目标对象的至少一部分的提示的一种。
如果在步骤S1021中，在所述图像中检测到所述边界框，则转到步骤S1022，通过检测所述边界框的边与所述图像的相应边之间是否有重合，来确定检测到的所述边界框是否为包含整个目标对象的完整边界框，进而确定是否能够检测到完整目标对象。
如果所述边界框的边与所述图像的相应边之间有重合(“是”)，则可确定检测到的所述边界框不是完整边界框(即不能够检测到完整目标对象)。此时，可转到步骤S105，提供检测不到完整边界框的提示。所述检测不到完整边界框的提示属于检测不到完整目标对象的提示的一种。
关于所述边界框的边与所述图像的相应边之间有重合，可以是所述边界框的至少一条边的至少一部分与所述图像的相应边的至少一部分之间有重合。即，不一定是整条边重合才属于重合，而是边的一部分出现重合(不是整条边重合)也属于重合。因此，不仅整条边重合属于检测不到完整边界框，边的一部分出现重合也属于检测不到完整边界框。
如上所述，在本公开中，还可以在确定检测到目标对象的边界框之后，进一步确定图像中检测到的目标对象的边界框是否为完整边界框，即，图像中是否包含完整的目标对象，或者说，图像中检测到的边界框是否能够包围整个目标对象。
例如，对于形状规则的身份证、读物(书、杂志等纸介质读物，或者电子书等电子读物)等的作为目标对象的物体，可以检测图像中拍到的该物体的轮廓是否露出。假如目标对象是一本纸介质的实体书，那么可以检测这本实体书的书边是否包含在该边界框中。例如，参照图2B，在该图像801’中检测到了完整边界框803’，该完整边界框803’包围住了目标对象802的各个实际边。对于形状不规则的目标对象，也可以进行类似的 检测，即确定图像中检测到的边界框是否包含完整的目标对象，例如可以通过检测到的边界框是否包含了该目标对象的轮廓来确定。
如上所述，在步骤S1022中，可以通过例如检测该边界框的边与该图像的相应边之间是否有重合，来确定检测到的边界框是否为包含整个目标对象的完整边界框，进而确定检测到的目标对象的至少一部分是否包含了整个目标对象。
如果所述边界框的边与所述图像的相应边至少部分重合(所述边之间的某个或某些距离等于0)，则可在步骤S105提供检测不到完整边界框的提示。
请注意，这里提供的是“检测不到完整边界框”的提示，与之前描述的“检测不到边界框”的提示不同。
如果所述边界框的边与所述图像的相应边之间没有重合(表示检测到完整边界框)，则可以转到步骤S104。
为了容易理解和方便描述起见，下面将以检测到的边界框为矩形框的情况为例，对此进行说明。
如图5A～5F所示，其中，图5A、5B是检测到的边界框的边与图像的相应边完全重合的情形，图5C、5D、5E、5F是检测到的边界框的边与图像的相应边部分重合的情形。
其中，所述提供检测不到完整边界框的提示包括提示调整目标对象与图像传感器之间的相对位置，使所述目标对象远离和/或偏移。
这里，提示使所述目标对象远离和/或偏移包括例如移动目标对象使其远离(放远)和/或偏移，比如可以相对于图像传感器的取景框来放远和/或进行偏移，或者移动图像传感器(例如，移动图像传感器本身，或者移动图像传感器所在设备例如可穿戴设备或视障阅读设备之类的设备以移动图像传感器)使图像传感器远离目标对象和/或偏移目标对象，也或者可以移动目标对象与图像传感器两者，以实现远离和/或偏移目标对象。即，调整目标对象与图像传感器之间的相对位置可以是调整目标对象的位置，也可以是调整图像传感器的位置，甚至也可以是两者的位置都进行调整。
例如，如图5A所示，如果检测到的矩形框与所述图像有四条边重合，说明目标对象有一部分超出了取景框的视野而导致目标对象没被采集全；或者如图5B所示，说明目标对象刚好在取景框的视野范围而被采集全了，但是目标对象在图像中过大，对于四条边重合的这种情况，可以直接提示使目标对象远离，以使得目标对象更多进入取景框的视野范围。
如果检测到的矩形框与所述图像有三条边出现重合或者两条边出现重合，如图5C、5D、5E、5F所示，说明目标对象可能没被采集全，则可以直接提示使目标对象远离(例如图5C中有三条边重合、图5D中有两条边重合)，和/或将所述目标对象沿与下述方位有关的方向移动：该方位通过所述边界框的、与所述图像的边重合的边所在的方位相反的方位来确定(例如图5E，该边界框的两条边与图像的两条边分别在左上方向重合，则可以沿与左上方向相反的右下方向移动目标对象)，或者通过沿所述与目标对象对应的边界框的中心位置相对于所述图像的中心位置或中心区域的方位相反的方位来确定(例如图5F，该边界框的两条边与图像的两条边分别在右上方向重合，边界框的中心位置相对于图像的中心位置偏右上方向，则可以沿与右上方向相反的左下方向移动目标对象)。
当检测到的边界框与所述图像的边框完全重合时，即，当检测到的边界框的各边与所述图像的相应边之间的距离均等于0时，在上面的例子中，将这种情况视为未检测到完整边界框。不过，对于检测到的边界框与图像的边框完全重合的这种特例情况，也可以视为检测到完整边界框，即，本公开并不局限于将这种情况视为未检测到完整边界框，而是可以视实际需求来决定是否将与图像的边框完全重合的边界框作为一个完整边界框。
对于图像识别来说，目标对象的边界框与图像的边框完全重合，可能会影响图像边缘附近的图像识别效果，因为目前的图像传感器所采集到的图像的边缘部分的清晰度相对低一些。所以，采集到的目标对象靠近图像中心时，识别效果会更好。因此，通过特定距离阈值的设置，可以使得获得的边界框小于图像的边框，由此可以使得所获得的边界框中的目标对象在所述图像中完整且尽量居中。另外，后面还将描述进一步通过其他方式或者进一步设定目标对象的边界框的边与图像的边的距离相关的距离范围参数来获得不仅完整、而且要大小合适且尽量居中的目标对象的一些示例性实施方式。
通过如上所述，自动地向用户给出针对各种相关情况的不同提示，能够有助于用户快速地调整目标对象与图像传感器之间的相对位置，更快地获得满足质量要求(例如目标对象在图像中的位置和/或大小等是否合适)的图像，从而有效地节约了处理资源，大大地缩短了处理时间。
根据本公开的一些实施例，如图1所示，在步骤S103提供调整所述目标对象至所述图像传感器视场范围内的提示之后，可以转到步骤S101，重新获取图像传感器重新采集的图像，并在步骤S102(更具体地，在图3中的步骤S1021)继续检测重新获取的图像中是否存在目标对象的至少一部分，直到在重新获取的图像中检测到目标对象的至少一部分。
在通过自动判断，例如确定“检测不到边界框”或者进一步地“检测不到完整边界框”的情况下，如上所述，根据一些实施例，将自动地向用户提供提示，以有助于用户快速地调整目标对象与图像传感器的相对位置，更快地获得满足质量要求(例如目标对象在图像中的位置和/或大小等是否合适)的图像，从而有效地节约了处理资源，大大地缩短了处理时间。
根据一些实施方式，本公开的目标检测方法和相关电路及设备等可以用于视障辅助和可穿戴设备等，在例如视力障碍用户或佩戴了不提供显示屏的可穿戴设备的用户无法判断图像质量的情况下，本公开的示例性实施方式可以通过自动帮助用户预先判断图像质量，并且通过进一步自动提示用户，获得更好质量的图像，以提高目标识别的准确度和速度，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
根据本公开的一些实施例，所述提供检测不到边界框的提示包括提示使所述目标对象远离和/或偏移。
根据一些实施例，检测不到边界框可能是因为目标对象不在图像传感器的采集范围(即视场范围)内，此时使目标对象远离(远离图像传感器)(通过移动目标对象和/或移动图像传感器)能够扩大采集范围，有可能把目标对象容纳进扩大后的采集范围。或者在当前的采集范围，可以偏移目标对象的位置，使其进入图像传感器的采集范围。再或者，用户可以移动目标对象使其远离图像传感器以扩大采集范围并且同时偏移目标对象使其进入图像传感器的采集范围。尤其对于例如不提供预览图的情况或者对于用户是视障人士的情况而言，给出这样清晰明确的提示是非常重要的，可以避免用户盲目的尝试，由此使得能够快速地采集到含有目标对象的图像。
如图1所示，如果在步骤S102，在所述图像中检测到目标对象的至少一部分，则可转到步骤S104。
在步骤S104，响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
这里的“几何关系”包括但不限于距离关系、面积比值关系、位置关系等等。
根据一些实施例，如前所述，在所述图像中检测到的所述目标对象的至少一部分可由在所述图像中检测到的对应的边界框来表示。相应地，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示包括：
如果所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
根据一些实施例，所述特定距离要求可以包括所述边界框的每条边与所述图像的相应边之间的距离均大于相应的特定距离阈值。
如图4所示，目标对象的矩形边界框与图像的四个边分别可以对应四个特定距离阈值，例如分别可以称为上侧距离阈值D up、下侧距离阈值D down、左侧距离阈值D left、右侧距离阈值D right(如图4中的虚线箭头所示)。只有当该边界框的每条边与该图像的相应边之间的距离均大于相应的特定距离阈值，即作为该边界框的上边与图像的上边之间的距离的上侧距离、作为该边界框的下边与图像的下边之间的距离的下侧距离、作为该边界框的左边与图像的左边之间的距离的左侧距离、作为该边界框的右边与图像的右边之间的距离的右侧距离(如图4中的实线箭头所示)分别大于与其相对应的上侧距离阈值D up、下侧距离阈值D down、左侧距离阈值D left、右侧距离阈值D right时，该边界框为完整边界框。
每个距离阈值的大小可以根据实际需求来设定，它们不一定相等。
根据本公开的一些实施例，每个距离阈值可以为0，也可以为大于0的值。由此，当检测到的边界框的边与所述图像的相应边的距离均大于各自相应的距离阈值时，检测到的该边界框为完整边界框，且该完整边界框小于图像的边框。
如上所述，每个距离阈值的大小可以根据实际需求来设定，本公开无需对其进行额外限制。不过，根据一些实施例，可以将各特定距离阈值设定为使得获得的完整边界框比图像边框小即可，即，无需将这里的特定距离阈值设置得过大。
顺便提及，当检测到的边界框的边与所述图像的相应边之间的距离有至少一个等于0时，出现前面描述过的边界框的边与图像的相应边有重合的情况，即前面描述的步骤S102(更具体地，步骤S1022)中检测边界框时判断边界框的边与图像的边之间是否有重合的情况。
根据一些实施例，所述如果所述边界框的每条边与所述图像的相应边之间的距离不符合特定距离要求，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示包括：
如果检测到的所述矩形框的每条边与所述图像的相应边之间的距离均小于相应的特定距离阈值，则直接提示使目标对象远离；以及
如果检测到的所述矩形框的边与所述图像的相应边之间的距离之中有三个或者更少数量的距离小于相应的特定距离阈值，则提示使目标对象远离，和/或使所述目标对象沿下述方位移动：该方位是沿所述边界框的中心位置相对于所述图像的中心位置或中心区域的方位相反的方位。
根据一些实施例，所述特定距离要求还可以包括所述边界框的边与所述图像的相应边之间的距离在相应的特定距离范围内。
由此，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
如果所述边界框的边与所述图像的相应边之间的距离不在相应的特定距离范围内，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
这里，通过进一步对边界框的边与图像的相应边之间的距离是否处于特定距离范围内的判断，来确定图像中的目标对象的位置是否合适。
虽然关于步骤S104所描述的实施例同样用到了目标对象的边界框的边与图像的相应边之间的距离，不过稍前的实施例通过例如检测目标对象的边界框的每条边与图像的相应边之间的距离是否均大于相应的特定距离阈值，来确定是否提供调整所述目标对象与所述图像传感器的相对位置的提示。而稍后的实施例是通过基于边界框的边与所述图像的相应边之间的距离是否在相应的特定距离范围内，确定是否提供调整所述目标对象与所述图像传感器的相对位置的提示。
换而言之，在步骤S104中，可以通过相应的特定距离阈值和/或相应的特定距离范围，来确定已经检测到的边界框的位置是否合适，以及如果不合适，可以例如确定该边界框的位置是否偏上、偏下、偏左、偏右等等(例如小于相应的特定距离阈值，或者小于相应的特定距离范围的下限或大于相应的特定距离范围的上限的情况等)。
另外，根据一些实施例，前述的与特定距离阈值有关的判断和上述的与特定距离范围有关的判断可以合并为一个判断步骤来实施。换言之，可以仅判断检测到的边界框的各边与图像的相应边之间的距离是否均在相应的特定距离范围内，便可以是进行了特定距离阈值判断和特定距离范围判断这两者，只要其中所述特定距离范围的下限为所述特定距离阈值即可。另一方面，当特定距离范围的下限大于所述特定距离阈值时，可以如上面描述的那样分开判断。总之，本公开的各实施例可以灵活变化，这里是否分开判断可以取决于实际的需求和/或对于处理资源的考虑等等。
总之，无论距离阈值还是距离范围，都是属于对于距离的要求，因此，本公开中可以统一使用“特定距离要求”来表述。关于特定距离要求，本公开包括但不限于上面例举出的距离阈值与距离范围。
通过上面的描述可知，无论是通过距离阈值还是通过距离范围来进行判断，能够符合相关距离要求的边界框都是完整边界框。
在本公开中，可以先进行边界框的判断，再进行其他条件的判断和/或调整位置的提示，以得到目标对象的位置和/或大小合适的图像。
也可以在进行边界框的判断之后，进一步进行是否完整边界框的判断，在判断是完整边界框的情况下，进一步进行其他的位置判断和/或调整位置的提示等操作，通常能更快地得到目标对象的位置和/或大小合适的图像。
图6将具体给出关于距离范围进行判断的例子。如图6所示，在检测到的边界框为矩形框的情况下，可以设定例如四个特定距离范围，分别称为上侧距离范围D up-range、下侧距离范围D down-range、左侧距离范围D left-range、右侧距离范围D right-range。
请注意，这里的距离范围与前面描述的“上侧距离阈值D up、下侧距离阈值D down、左侧距离阈值D left、右侧距离阈值D right”不同。主要的不同在于，每个距离范围通过各自的距离范围下限和距离范围上限这两者来限定，而前述的特定距离阈值是一个值。进一步地，每个距离范围的下限可以大于或等于上述的相应的距离阈值，在上述的合并判断的情况下，距离范围的下限等于相应的距离阈值，而在分开判断的情况下，距离范围的下限大于相应的距离阈值。这里，特定距离范围的判断条件比特定距离阈值的判断条件相对严格。
在图6中，上侧、下侧、左侧、右侧距离范围分别由图中的虚线箭头示出，由上侧、下侧、左侧、右侧距离范围的下限构成图6中的大虚线框，由上侧、下侧、左侧、右侧距离范围的上限构成图6中的小虚线框(为了清楚起见，图6中未示出边界框内的目标对象)。如果检测到的边界框的各边都位于该大虚线框与该小虚线框之间，如图6中的边界框803在该大虚线框与该小虚线框之间，说明该边界框的各边与所述图像的相应边之间的距离在各自相应的特定距离范围内(该边界框的各边与所述图像的相应边之间的距离均大于各自距离范围的下限且小于各自距离范围的上限)，则可能无需提示调整目标对象与图像传感器之间的相对位置，或者还可以通过其他条件来进一步确定是否需要提示调整该相对位置。否则，如果检测到的边界框的各边中至少一个边不位于该大虚线 框与该小虚线框之间(该边界框的各边与所述图像的相应边之间的距离中至少一个距离不在相应的特定距离范围内)，则认为可以提示调整该相对位置。
这里，如上所述，通过进一步对边界框的边与图像的相应边之间的距离是否处于特定距离范围内的判断，即，既有对于所述的边之间的距离的上限控制，又有对于所述的边之间的距离的下限控制，来确定图像中的目标对象的位置和/或大小是否合适。例如，判断所述的边之间的距离是否处于一定的距离范围内，能够进一步反映图像质量，比如当所述的边之间的某个或某些距离小于相应的距离范围的下限时，目标对象可能太靠近图像的边缘，由于图像边缘的分辨率低于图像中心部分的分辨率，因此可以通过提示用户如何调整目标对象与图像传感器之间的相对距离(将在后面进行描述)，以使得目标对象离开取景框边缘，靠近取景框中心，从而能够获得更好质量的图像。再比如当所述的边之间的某个或某些距离大于相应的距离范围上限时，可能目标对象在图像中的位置有偏移或者整个目标对象在图像中的尺寸过小，影响图像的质量，乃至进一步影响向用户提供的服务(比如视障阅读服务、引导服务等)，由此同样可以通过提示用户如何调整目标对象与图像传感器之间的相对距离(将在后面进行描述)，以使得目标对象在图像中的位置和/或大小更合适，从而能够获得更好质量的图像。
根据一些实施例，步骤S104中的所述如果所述边界框的边与所述图像的相应边之间的距离不在相应的特定距离范围内，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
分别检测所述边界框的上、下、左、右边与所述图像对应的上、下、左、右边之间的距离，相应地得到所述边界框与所述图像之间的上侧距离、下侧距离、左侧距离、右侧距离，即得到作为该边界框的上边与图像的上边之间的距离的上侧距离、作为该边界框的下边与图像的下边之间的距离的下侧距离、作为该边界框的左边与图像的左边之间的距离的左侧距离、作为该边界框的右边与图像的右边之间的距离的右侧距离(由于篇幅所限，图6中未示出这四个距离)；以及
对于与所述图像的上、下、左、右边分别对应的上侧距离范围、下侧距离范围、左侧距离范围、右侧距离范围，每一个所述距离范围均由各自的距离上限和距离下限来共同限定，如果所述上侧距离、下侧距离、左侧距离、右侧距离中至少一个不在各自对应的上侧距离范围、下侧距离范围、左侧距离范围、右侧距离范围内(即，上侧距离、下侧距离、左侧距离、右侧距离中至少一个不在图6所示的大虚线框与小虚线框之间的位置)，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
请注意，可以根据实际需求来设定上述的特定距离范围(如图6中所示，上侧、下侧、左侧、右侧距离范围下限与上侧、下侧、左侧、右侧距离范围上限)，本公开无需对此进行限制。举一个例子，对于普通图像传感器采集的普通图像，可以设定上侧、下侧、左侧、右侧的距离范围均为[图像相应边的长度对应的像素数的10％,图像相应边的长度对应的像素数的40％]。总之，所述距离范围下限可以为0或者大于0的值，而所述距离范围上限应为大于相应的距离范围下限的值。另外，上侧、下侧、左侧、右侧的特定距离范围可以不相等。当然，也可以如这里的例子一样，四个特定距离范围都相等。
请注意，图6中示出的仅是一个用于说明如何基于边界框的边与所述图像的相应边之间的距离是否在相应的特定距离范围内，确定是否提示调整所述目标对象与所述图像传感器之间的相对位置的示例。本领域技术人员可知，本公开不限于该示例。
根据一些实施例，步骤S104中的响应于所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，提供调整所述目标对象与所述图像传感器之间的相对位置的提示包括：
如果所述边界框的边与所述图像的相应边之间的距离中至少一个小于相应的特定距离范围的下限，则提示使目标对象远离和/或沿所述边界框的小于所述相应的特定距离范围下限的边所在方位相反的方向使目标对象偏移；以及
如果所述边界框的边与所述图像的相应边之间的距离中至少一个大于相应的特定距离范围的上限，则提示使目标对象靠近和/或沿与所述边界框的大于所述相应的特定距离范围的上限的边所在方位一致的方向使目标对象偏移。
如图7A、7B、7C所示，在检测到的边界框的边与所述图像的相应边之间的距离中至少一个小于相应的特定距离范围下限的情况下，提示用户使目标对象远离(参考图7A)和/或沿所述边界框的小于所述相应的特定距离范围下限的边所在方位相反的方向使目标对象偏移(参考图7C)；在检测到的边界框的边与所述图像的相应边之间的距离中至少一个大于相应的特定距离范围上限的情况下，提示使目标对象靠近(参考图7B)和/或沿与所述边界框的大于所述相应的特定距离范围上限的边所在方位一致的方向使目标对象偏移(参考图7C)。
这里，如上所述，当判断边界框的边与图像的相应边之间的距离不处于特定距离范围内时，确定需要改善图像质量，在本实施方式中，通过提示用户如何调整目标对象与图像传感器之间的相对距离，使得目标对象在图像中的位置和/或大小更合适，从而能够获得更好质量的图像。
请注意，图7A、7B、7C给出了“检测到的边界框的边与所述图像的相应边之间的距离不在相应的特定距离范围内”的几种示例情况以便容易理解，不代表本公开仅限于这几种示例情况。
另外，根据一些实施例，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
如果所述边界框的中心位置不在所述图像的中心区域中，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
这里，通过边界框的中心位置与图像的中心区域的比较，来判断图像中采集到的目标对象的位置是否居中。
根据一些实施例，所述边界框的中心位置可以通过该边界框在图像中的位置信息来确定。具体地例如，所述边界框的中心位置可以通过所述边界框在图像中的各顶点位置来确定。除了顶点位置之外，还可以通过边界框在图像中的其他位置信息来确定边界框的中心位置，本公开对此不作具体的限制。
另外，所述图像的中心区域可以是围绕该图像的中心点的一个区域。可以根据实际情况来确定中心区域的大小和形状，本公开对于其大小和形状不做特别的限制。对于中心区域的形状，例如可以是矩形、圆形等。
这里的中心点可以是例如几何中心，所以，图像的中心点例如也可以是中心区域的中心点(几何中心)。
这里，如上所述，通过对边界框的中心位置是否处于所述图像的中心区域中的判断，来确定图像中的目标对象的位置是否居中。例如，判断边界框的中心位置是否处于所述图像的中心区域中，能够进一步反映图像质量，比如当边界框的中心位置不处于所述图像的中心区域中时，目标对象可能位置没有居中，有些偏离图像的中心区域(或者中心位置)，影响图像的质量，乃至进一步影响向用户提供的服务(比如视障阅读服务、引导服务等)，由此同样可以通过提示用户如何调整目标对象与图像传感器之间的相对距离(将在后面进行描述)，以使得目标对象在图像中的位置更合适(比如尽量居中)，从而能够获得更好质量的图像。
图8A示出了检测到的目标对象的边界框803的中心位置805不在所述图像801的中心区域804中的情形，此时可以提供调整所述目标对象与所述图像传感器之间的相对位 置的提示。另外，图8B示出了检测到的目标对象的边界框803的中心位置805在所述图像801的中心区域804中的情形，此时可以根据需要进行下一步操作。
根据一些实施例，步骤S104中的所述如果所述边界框的中心位置不在所述图像的中心区域中，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
基于所述边界框的中心位置相对于所述图像的中心区域的方向，提示用户沿与该方向相反的方向移动所述目标对象。
如前面描述了图8A中，检测到的边界框803的中心位置805不在所述图像801的中心区域804中的情形。从图8A可知，中心位置805在中心区域804的偏右上方向，则可以确定：可以沿与右上方向相反的左下方向移动目标对象，以使边界框803的中心位置805进入图像801的中心区域804中。
这里，如上所述，当判断边界框的中心位置不在所述图像的中心区域中时，确定需要改善图像质量，在本实施方式中，通过提示用户如何调整目标对象与图像传感器之间的相对距离，使得目标对象在图像中的位置更合适，从而能够获得更好质量的图像。
另外，例如，根据一些实施例，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
如果所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
这里，通过面积比值的判断，来确定图像中采集到的目标对象的大小是否合适。
如图9A所示，边界框的面积与图像的面积的比值大约为1:9，面积比值过小；如图9B所示，边界框的面积与图像的面积的比值大约为94:100，比值过大；如图9C所示，边界框的面积与图像的面积的比值大约为3:4，面积比值合适。
关于上述的特定比值范围，可以根据具体的需求来进行设定，比如50％～85％的比值范围，或者其他的范围等等，本公开不作限制。
这里，如上所述，通过对边界框的面积与图像的面积的比值是否在特定比值范围内的判断，来确定图像中的目标对象的大小是否合适。例如，判断边界框的面积与图像的面积的比值是否处于特定比值范围(符合该比值范围的图像，图像大小较为合适，图像质量较好)中，能够进一步反映图像质量，比如当边界框的面积与图像的面积的比值不处于特定比值范围时，目标对象在图像中可能过大或过小，影响图像的质量，乃至进一步影响向用户提供的服务(比如视障阅读服务、引导服务等)，由此同样可以通过提示 用户如何调整目标对象与图像传感器之间的相对距离(将在后面进行描述)，以使得目标对象在图像中的大小更合适，从而能够获得更好质量的图像。
根据一些实施例，步骤S104中的所述如果所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
如果所述边界框的面积与所述图像的面积的比值高于所述特定比值范围的上限，则提示使目标对象远离；以及
如果所述边界框的面积与所述图像的面积的比值低于所述特定比值范围的下限，则提示使目标对象靠近。
例如，对于如图9A所示的面积比值过小的情形，可以提示使目标对象靠近。反之，对于如图9B所示的面积比值过大的情形，可以提示使把目标对象远离。
这里，如上所述，当判断边界框的面积与所述图像的面积的比值不在所述特定比值范围中时，确定需要改善图像质量，在本实施方式中，通过提示用户如何调整目标对象与图像传感器之间的相对距离，使得目标对象在图像中的大小更合适，从而能够获得更好质量的图像。
根据一些实施例，所述目标对象中包含文本区域，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
确定所述目标对象的文本区域的文本行相对于所述图像的一条边的角度(以下简称为“文本行的倾斜角度”)是否位于特定角度范围内，如果所述角度不位于所述特定角度范围内，则提示用户调整所述目标对象与所述图像传感器之间的相对角度。
对于上述的特定角度范围，例如可以为[-30°，30°]。考虑到算法精度以及采集稳定性等因素，本公开中，在文本行的倾斜角度位于所述特定角度范围内的情况下，即认为文本行基本上为水平的。可以理解的是，在算法更精细化以及采集稳定性得到提高的情况下，可以合理调整所述特定角度范围，例如可以调整为：[-20°，20°]、[-10°，10°]等等。
在本公开中，文本行的倾斜角度可以指文本行(比如可以选取第一个文本行)相对于所述图像的某条边(为了便于理解和描述，可以将该边定义为水平边，所述图像的与该边相邻的另一条边定义为垂直边)的角度。“文本行水平”可以指文本行基本与所述图像的水平边平行(例如，文本行的倾斜角度小于30度)。如果超出所述特定角度范围， 则可以认为“文本行倾斜”，例如，图10中的文本行806与所述图像的水平边之间形成了大约45度的角度。在“文本行倾斜”的情况下，可以提示用户调整所述目标对象与所述图像传感器之间的相对角度。
根据一些实施例，可以通过以下方法来检测文本行的倾斜角度：基于Hough变换的方法、基于交叉相关性的方法、基于投影的方法、基于Fourier变换的方法和K-最近邻簇方法等。
根据一些实施例，可以通过调整目标对象的角度来调整目标对象与图像传感器之间的相对角度，所提供的目标对象的相对角度提示信息可以包括但并不局限于目标对象的旋转方向及旋转角度，也可以仅包括旋转方向或旋转角度，另外还可以包括其它提示信息，例如可以播放“请调整目标对象的角度”之类的语音。
这里，如上所述，通过对图像中目标对象的文本区域的文本行的倾斜角度的判断，来确定图像中的目标对象的位置和/或姿态是否合适。例如，判断图像中目标对象的文本区域的文本行的倾斜角度是否处于特定角度范围内，能够进一步反映图像质量，比如当图像中目标对象的文本区域的文本行的倾斜角度不在所述特定角度范围内时，影响对于图像中的文本的识别速度和准确度，乃至进一步影响向用户提供的服务(比如视障阅读服务、引导服务等)，此时可能需要调整目标对象的姿态。因此，这里通过提示用户如何调整目标对象与图像传感器之间的相对角度等，使得目标对象在图像中的位置和/或姿态更合适，从而能够获得更好质量的图像。
在上述的文本行的倾斜角度大于所述特定角度范围的情况下，可以使得图像传感器重新采集图像并继续检测，直到重新采集的图像中的文本行的倾斜角度不小于所述特定角度范围，以获得满足质量要求的图像。
上面给出了三种基于边界框与图像的几何关系来判断是否提示用户调整目标对象与图像传感器之间的相对位置的示例，即，边界框的边与图像的边的距离(特定距离要求，比如可以包括基于距离阈值的距离要求以及基于距离范围的距离要求)、边界框的中心位置是否在图像的中心区域、边界框的面积与图像的面积的比值。另外，还给出了一种基于文本行的倾斜角度来判断是否提示用户调整目标对象与图像传感器之间的相对角度的示例。请注意，本公开不限于上面示例的这四种判断条件，而是还可以使用其他的判断条件，考虑到篇幅，在此不再展开叙述。
另外，上面给出的三种基于边界框与图像的几何关系的判断条件可以如上所述那样分别进行判断，也可以对这些条件进行任意组合来进行判断。比如，可以将其两两组合， 例如先判断边界框的中心位置是否在图像的中心区域，再判断面积比值是否合适；当然，也可以反过来，先判断面积比值是否合适，再判断中心位置是否合适。也可以三个条件互相组合，比如先判断边界框的中心位置是否在图像的中心区域，再判断面积比值是否合适，最后可以判断边的距离是否符合特定距离要求(例如在设定的距离阈值之上和/或在设定的距离范围内)等等。另外，也可以将文本行的倾斜角度的判断与其他条件进行组合。判断条件越多，得到的目标对象在图像中的位置、大小等等可能越符合要求，但是相应地需要的计算资源和时间也会稍多。
通过组合判断，能够得到质量更好的图像，自动向用户提供更准确的提示，可以大大缩短处理时间，有效节约处理资源，而且还能够向用户提供更好更精准的服务。
为了更清楚起见，下面将结合示例来描述两两组合的方式的判断过程。如前所述，所述目标对象的至少一部分可由在所述图像中检测到的对应的边界框来表示。
根据一些实施例，如图11A所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的边与所述图像的相应边之间的距离是否符合特定距离要求；
如果所述边界框的边与所述图像的相应边之间的距离符合特定距离要求，则转到步骤S1042，确定所述边界框的中心位置是否处于所述图像的中心区域中；以及
如果确定所述边界框的中心位置不处于所述图像的中心区域中，则转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
另外，如果所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，则可以转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
这里，所述特定距离要求可以包括例如上述的基于距离阈值的距离要求和/或基于距离范围的距离要求。
本实施方式是将边界框与图像的边的距离的判断条件与边界框的中心位置的判断条件组合在一起，用于判断目标对象在图像中的位置和/或大小是否合适。在通过自动判断，确定不符合要求的情况下，如上所述，根据一些实施例，将自动地向用户提供提示，以有助于用户快速地调整目标对象与图像传感器的相对位置，更快地获得满足质量要求(例 如目标对象在图像中的位置和/或大小等是否合适)的图像，从而有效地节约了处理资源，大大地缩短了处理时间。
根据一些实施方式，在例如视力障碍用户或佩戴了不提供显示屏的可穿戴设备的用户无法判断图像质量的情况下，本公开的示例性实施方式通过自动帮助用户预先判断图像质量，并且通过进一步自动提示用户，获得更好质量的图像，以提高目标识别的准确度和速度，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
通过自动帮助用户预先判断图像质量，并且通过进一步自动提示用户，获得更好质量的图像，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
根据一些实施例，如图11B所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的边与所述图像的相应边之间的距离是否符合特定距离要求；
如果所述边界框的边与所述图像的相应边之间的距离符合特定距离要求，则转到步骤S1042，确定所述边界框的面积与所述图像的面积的比值是否在特定比值范围内；以及
如果确定所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则可以转到步骤S1043，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
另外，如果所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，则可以转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式是将边界框与图像的边的距离的判断条件与面积比值的判断条件组合在一起，用于判断目标对象在图像中的位置和大小是否合适。与上面类似地，在通过自动判断，确定不符合要求的情况下，如上所述，根据一些实施例，将自动地向用户提供提示，以有助于用户快速地调整目标对象与图像传感器的相对位置，更快地获得满足质量要求(例如目标对象在图像中的位置和大小等合适)的图像，从而有效地节约了处理资源，大大地缩短了处理时间。同样，在例如视力障碍用户或佩戴了不提供显示屏的可穿戴设备的用户无法判断图像质量的情况下，通过自动帮助用户预先判断图像质量，并且 通过进一步自动提示用户，获得更好质量的图像，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
根据一些实施例，如图11C所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的中心位置是否处于所述图像的中心区域中；
如果确定所述边界框的中心位置处于所述图像的中心区域，则转到步骤S1042，确定所述边界框的面积与所述图像的面积的比值是否在特定比值范围内；以及
如果确定所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则转到步骤S1043，以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
另外，如果所述边界框的中心位置不处于所述图像的中心区域中，则可以转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式是将边界框的中心位置的判断条件与面积比值的判断条件组合在一起，用于判断目标对象在图像中的位置和大小是否合适。与上面类似地，在通过自动判断，确定不符合要求的情况下，将自动地向用户提供提示，以有助于用户快速地调整目标对象与图像传感器的相对位置，更快地获得满足质量要求的图像，从而有效地节约了处理资源，大大地缩短了处理时间。同样，在例如视力障碍用户或佩戴了不提供显示屏的可穿戴设备的用户无法判断图像质量的情况下，通过自动帮助用户预先判断图像质量，并且通过进一步自动提示用户，获得更好质量的图像，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。后面几种组合方式的有益技术效果与此类似，下面将不再赘述。
根据一些实施例，如图11D所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的中心位置是否处于所述图像的中心区域中；
如果确定所述边界框的中心位置处于所述图像的中心区域中，则转到步骤S1042，确定所述边界框的边与所述图像的相应边之间的距离是否符合特定距离要求；以及
如果确定所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，则转到步骤S1043，以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
同样，如果所述边界框的中心位置不处于所述图像的中心区域中，则可以转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式是将边界框的中心位置的判断条件与所述边界框与图像的边的距离的判断条件换了顺序组合在一起，用于判断目标对象在图像中的位置是否合适。
根据一些实施例，如图11E所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的面积与所述图像的面积的比值是否在特定比值范围内；
如果确定所述边界框的面积与所述图像的面积的比值在特定比值范围内，则转到步骤S1042，确定所述边界框的边与所述图像的相应边之间的距离是否在特定距离范围内；以及
如果确定所述边界框的边与所述图像的相应边之间的距离不在特定距离范围内，则转到步骤S1043，以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
另外，如果所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则可以转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式是将面积比值的判断条件与边界框的中心位置的判断条件换了顺序组合在一起，用于判断目标对象在图像中的位置和大小是否合适。
根据一些实施例，如图11F所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的面积与所述图像的面积的比值是否在特定比值范围内；
如果确定所述边界框的面积与所述图像的面积的比值在特定比值范围内，则转到步骤S1042，确定所述边界框的中心位置是否处于所述图像的中心区域中；以及
如果确定所述边界框的中心位置不处于所述图像的中心区域中，则转到步骤S1043，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
另外，如果所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则可以转到步骤S1043以提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式是将面积比值的判断条件与边界框的中心位置的判断条件换了顺序组合在一起，用于判断目标对象在图像中的位置是否合适。
下面将举例说明三个判断条件组合的实施方式。通过组合更多的判断方式，能够得到质量更好的图像，自动向用户提供更准确的提示，可以大大缩短处理时间，有效节约处理资源，而且还能够向用户提供更好更精准的服务。
根据一些实施例，如图12A所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述边界框的边与所述图像的相应边之间的距离是否符合特定距离要求；
如果所述边界框的边与所述图像的相应边之间的距离符合特定距离要求，则转到步骤S1042，确定所述边界框的中心位置是否处于所述图像的中心区域；
如果确定所述边界框的中心位置处于所述图像的中心区域，则转到步骤S1043，确定所述边界框的面积与所述图像的面积的比值是否在特定距离范围内；以及
如果确定所述边界框的面积与所述图像的面积的比值不在特定比值范围内，则转到步骤S1044，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式通过将边的距离、中心位置、面积比值三者按顺序组合在一起，用于判断目标对象在图像中的位置和大小是否合适，以便得到在图像中位置和大小更合适的目标对象。
根据一些实施例，如图12B所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述目标对象的所述中心位置是否处于所述图像的中心区域；
如果所述中心位置处于所述中心区域，则转到步骤S1042，确定所述边界框的面积与所述图像的面积的比值是否在特定距离范围内；
如果所述边界框的面积与所述图像的面积的比值在特定距离范围内，则转到步骤S1043，确定所述边界框的各边与所述图像的相应边的距离是否符合特定距离要求；以及
如果所述边界框的各边与所述图像的相应边的距离不符合特定距离要求，则转到步骤S1044，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式通过将面积比值、边的距离、中心位置三者按顺序组合在一起，同样用于判断目标对象在图像中的位置和大小是否合适，以便得到在图像中位置和大小更合适的目标对象。
根据一些实施例，如图12C所示，步骤S104中的所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示可以包括：
步骤S1041，确定所述中心位置是否处于所述图像的中心区域；
如果确定所述中心位置处于所述图像的中心区域，则转到步骤S1042，确定所述边界框的面积与所述图像的面积的比值是否在特定比值范围内；
如果确定所述边界框的面积与所述图像的面积的比值在特定比值范围内，则转到步骤S1043，确定所述边界框的各边与所述图像的相应边的距离是否符合特定距离要求；以及
如果确定所述边界框的各边与所述图像的相应边的距离不符合特定距离要求，则可以转到步骤S1044，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
本实施方式通过将中心位置、面积比值、边的距离三者按顺序组合在一起，同样用于判断目标对象在图像中的位置和大小是否合适，以便得到在图像中位置和大小更合适的目标对象。而且，相对来说，这种组合方式能够更快速得到图像中位置和大小都合适的目标对象。
另外，如上所述，文本行的倾斜角度的判断条件也可以与其他条件进行组合，下面将结合示例来简单描述。
如图12D所示，所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示还可以包括：
当所述目标对象中包含文本区域时，在步骤S1041’，确定所述目标对象的文本区域的文本行相对于所述图像的一条边的角度是否位于特定角度范围内；
如果所述目标对象的文本区域的文本行相对于所述图像的一条边的角度不位于特定角度范围内，则可以在步骤S1042’提示用户调整所述目标对象与所述图像传感器之间的相对角度。
如前所述，所述特定角度范围例如可以为[-30°，30°]，所述相对角度例如可以包括旋转角度和旋转方向这两个参量。
本实施方式给出了一种通过文本行倾斜角度来判断目标对象在图像中的位置是否合适的示例，这种方式同样可以如前面所述的那样，与例如中心位置、面积比值、边的距离这三者之中的至少一种进行组合，用于判断目标对象在图像中的位置和/或大小是否合适。因此，图11D除了步骤S1041’、S1042’之外，还可以包括前述的图10A～10F、图11A～11C中示出的那些步骤。换而言之，步骤S1041’、S1042’可以分别与图10A～10F、图11A～11C中的步骤进行组合。限于篇幅，在此不再展开叙述。
判断条件的组合方式不限于上面例举的那些，而且，如上所述，判断条件也不止上面描述的这几种。关于判断条件及其组合方式，本公开例举了足够多的示例来使得本公开更容易理解。而且，本公开不仅应当包括示例出的这些实施方式，还应该包括各种各样的替换方式、变形方式、以及扩展方式。
上面对于步骤S104进行了详细描述，下面将进一步描述其它步骤。
如前所述，在步骤S103提供调整所述目标对象至所述图像传感器视场范围内的提示之后，如图1所示，可以转到步骤S101，重新获取图像传感器重新采集的图像并在步骤S102继续检测重新获取的图像中是否存在目标对象的至少一部分，直到在重新获取的图像中检测到目标对象的至少一部分。
此外，在步骤S104提供调整所述目标对象与所述图像传感器之间的相对位置的提示之后，如图1所示，也可以转到步骤S101，重新获取图像传感器重新采集的图像并基于重新获取的图像继续执行步骤S102、S104等操作，直至在重新获取的图像中检测到目标对象的至少一部分，且确定该检测到的目标对象的至少一部分与该重新获取的图像的几何关系不再满足需要提供调整所述目标对象与所述图像传感器之间的相对位置的提示的特定条件、或满足无需提供调整所述目标对象与所述图像传感器之间的相对位置的提示的条件(即目标对象在图像中的大小和/或位置等已较为合适，无需再提供调整所述目标对象与所述图像传感器之间的相对位置的提示)。
换言之，对于检测不到目标对象或者检测到的目标对象在图像中的大小和/或位置等不合适的情形，均可以重新采集图像，以检测重新采集的图像中是否存在目标对象以及存在的目标对象的大小和/或位置是否合适，等等。
另外，根据一些实施例，可以控制重新采集和检测以及提示的次数和/或时间。例如，可以由用户主动(人为地)启动停止检测指示；或者由处理器等设备自行启动停止检测指示，比如可以基于检测的次数、连续检测不到边界框的时间等，来停止图像的继续获取、继续检测、以及停止因为检测不到边界框或完整边界框、或者因为目标对象与图像 传感器的相对位置持续不合适等等而发出的提示，从而避免过于集中和频繁的提示，实现向用户提供满足需求的服务。
根据一些实施例，可以使得图像传感器在至少间隔特定时间(比如2s)之后，才重新采集目标对象，以使得有足够的时间根据提示来移动目标对象和/或图像传感器所在设备，防止在还没有准备好目标对象与图像传感器的相对位置的情况下，过于频繁地采集目标对象。
使图像传感器间隔特定时间再重新采集目标对象的实现方式例如可以采用间隔采集，即，以固定时间为间隔来进行采集，比如每2秒一次快门，在某些图像传感器中可以设定这种间隔采集方式。
根据一些实施例，如图13所示，可以在使得图像传感器执行图像的采集之前，在步骤S110，通过检测指引物，确认要采集的目标对象，以便在采集之前，先确认目标对象，防止图像中出现多个疑似目标对象，给后续的检测和识别造成困难。
根据一些实施例，所述指引物例如可以为用户的手指。
这里，可以使用例如计算机视觉算法来检测手指并确定手指的位置，例如，可以采用来自OpenCV的算法，比如CascadeClassifier函数(https://docs.opencv.org/3.2.0/d1/de5/classcv_1_1CascadeClassifier.html)，从图像中确定手指的位置，从而确认目标对象。
另外，所述指引物并不局限于手指，例如还可以是笔。
根据一些实施例，在没有确认过目标对象却在图像中检测到多个疑似目标对象的情况下，还可以通过文字或声音提示的形式，提示用户确认要采集的目标对象。
另外，根据一些实施例，如图13所示，本公开的目标检测方法还可以包括：步骤S106，识别所述图像中的所述目标对象并输出识别的目标对象。例如，响应于在图像中检测到目标对象的至少一部分，且确定该检测到的目标对象的至少一部分与该图像的几何关系不满足需要提供调整所述目标对象与所述图像传感器之间的相对位置的提示的特定条件、或满足无需提供调整所述目标对象与所述图像传感器之间的相对位置的提示的条件(即目标对象在图像中的大小和/或位置等较为合适，无需提供调整所述目标对象与所述图像传感器之间的相对位置的提示)，可识别所述图像中的所述目标对象并输出识别的目标对象。
根据一些实施例，在提示用户移动目标对象的场景中，如上所述，所述目标对象可以是可以移动的目标对象，例如身份证、名片、护照、驾照、读物(媒介读物或者电子读物)、平板计算机、移动电话等等。
根据一些实施例，所述识别图像中的目标对象可以包括识别目标对象中的文字，相应地，所述输出识别的目标对象可以包括输出所识别的目标对象中的文字。
根据一些实施例，输出的形式包括声音形式的输出和/或文字形式的输出等。
在例如视力障碍用户或佩戴了不提供预览图像的可穿戴设备的用户无法判断图像质量的情况下，本公开的示例性实施方式通过自动帮助用户预先判断图像质量，并且通过进一步自动提示用户，获得更好质量的图像，以提高目标识别的准确度和速度，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
根据一些实施例，可以使得图像传感器先采集预览图像以便基于预览图像检测目标对象，并且在检测到目标对象的边界框或者检测到目标对象的完整边界框之后，再拍摄用于识别的图像(即拍照图像)，其中，所述预览图像的分辨率小于所述用于识别的图像的分辨率。
这样，通过在检测时使用低分辨率的图像，在识别时使用高分辨率的图像，使得在拍摄位置和/或大小合适的目标对象之前，采集的都是低分辨的图像，大大节约了有限的处理资源，而在真正识别时使用的是高分辨图像，由此又能够保证识别效果。
根据一些实施例，还可以在图像传感器采集或确认目标对象与图像传感器之间的相对位置合适从而要拍摄目标对象之前，提供保持目标对象的提示，以使得用户保持目标对象的位置，或者如果用户需要手持或扶持目标对象，可以使用户提前做好准备，手持或扶持得稳一些，以便于拍摄到清晰的图像。
根据一些实施例，上述的提示可以包括声音提示、震动提示、文字提示、图像提示、视频提示中的一种或者它们的任意组合。也就是说，本公开不限于某一种提示方式，而是可以多种方式组合起来提示。
对于通过震动来进行提示，比如提示远近，可以通过不同的震动模式来进行。比如短促的震动模式是需要放近，长而慢的震动模式表示需要放远等等。另外，关于通过震动来提示移动方向，例如可以通过相应方向上的震动来表明应该往该方向移动。震动提示对于无法听到正常声音的听障人士非常方便。
上面结合附图，描述了本公开的目标检测方法。本公开针对不提供预览图或者无法看到预览图的情形，提出了先在图像传感器采集的图像中检测是否存在目标对象的至少一部分，然后在未检测到目标对象的至少一部分的情况下，提供调整所述目标对象至所述图像传感器视场范围内的提示，或者，在检测到目标对象的至少一部分的情况下，基于目标对象在所采集的图像中的位置信息，提供调整目标对象与图像传感器的相对位置的提示，以便更方便地调整目标对象和/或图像传感器的位置，从而使得图像传感器更快地采集到位置和/或大小合适的目标对象。
由此，在例如视力障碍用户或佩戴了不提供预览图像的可穿戴设备的用户无法判断图像质量的情况下，本公开的示例性实施方式通过自动帮助用户预先判断图像质量，并且通过进一步自动提示用户，获得更好质量的图像，以提高目标识别的准确度和速度，从而有效地节约了处理资源，大大地缩短了处理时间，帮助用户更快地获得所需的服务(例如阅读服务、引导服务等等)。
根据本公开的另一方面，还提供一种电子电路，可以包括：被配置为执行上述的方法的步骤的电路。
根据本公开的另一方面，还提供一种视障辅助设备，包括：图像传感器，被配置为采集图像；上述的电子电路，所述电子电路例如被配置为检测所述图像中是否存在目标对象的至少一部分，以及响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。
根据一些实施例，所述电子电路例如还可被配置为响应于所述图像传感器为位置固定的设备，提供移动所述目标对象以使所述目标对象进入所述图像传感器视场范围内的提示；以及响应于所述图像传感器为位置可调的设备，提供移动所述目标对象和/或所述图像传感器以使所述目标对象进入所述图像传感器视场范围内的提示。
根据一些实施例，所述视障辅助设备可包括下述设备中的一种或多种：可连接至可穿戴设备(例如眼镜、头盔等)的视障辅助设备(例如阅读辅助设备等)、手持式视障辅助设备以及桌面式视障辅助设备，等等。
根据一些实施例，所述视障辅助设备可安装于相应的支架例如桌面式支架上以作为桌面式视障辅助设备。换言之，所述视障辅助设备还可包括相应的支架。
示例的，图14A和图14B分别示出了根据本公开的示例性实施例的用于视障辅助设备的、处于展开状态和折叠状态的支架例如桌面式支架的结构框图。如图14A可知，所述支架可包括连接件141、伸缩立臂142、底座旋转关节143以及带配重底座144。
所述连接件141被配置为安装(例如通过吸附安装)所述图像传感器以及所述电子电路。根据一些实施例，所述连接件141可包括磁铁，所述磁铁嵌入形成在所述连接件141的表面(例如，当所述支架以展开状态稳固于桌面时，所述连接件141的面向桌面的表面)以用于将所述图像传感器以及所述电子电路吸附安装在所述连接件141上。
所述伸缩立臂142的一端与所述连接件141相连，另一端通过所述底座旋转关节143与所述带配重底座144相连，被配置为支撑所述连接件141并通过伸缩和/旋转调节所述连接件141相对于所述带配重底座144或桌面的位置(包括高度、角度等等)。根据一些实施例，所述伸缩立臂142可包括相互连接的多个伸缩节，例如，如图14A所示出的4个伸缩节，当然，根据实际需求也可包括其它数量的伸缩节。所述多个伸缩节中的至少一个伸缩节的长度可调。另外，所述多个伸缩节中的与底座旋转关节143相连的底端伸缩节相对所述带配重底座144位置可调(例如，可相对带配重底座144在设定角度范围内旋转，所述设定角度范围例如可为包含端值的0～90°等)。其它伸缩节可套接于与其相连的前一伸缩节，或者可相对与其相连的前一伸缩节位置可调(例如，可相对前一伸缩节在设定角度范围内旋转，所述设定角度范围例如可为包含端值的0～90°等)。
所述带配重底座144被配置为支撑所述支架的其它各组件例如连接件141和伸缩立臂142，以使得所述支架处于展开状态(或工作状态)时能够稳固于桌面。例如，当所述支架的各组件处于图14A所示的状态时，所述带配重底座144能够使得所述支架稳固于桌面。根据一些实施例，如图14B所示，所述带配重底座144的横截面可为四端呈圆弧状的近似矩形形状，其中，矩形的长边的尺寸可为150～180mm，例如图中示出的175mm等，矩形的短边的尺寸可为60～90mm，例如图中示出的80mm等等。当然，根据实际需求，所述带配重底座144的横截面也可被设置为其它形状，例如圆形、椭圆形等等。
根据一些实施例，如图14A以及图14B所示，所述底座旋转关节143可位于所述带配重底座144的一侧的边缘，一方面，当所述支架处于展开状态时可以起到更好的支撑作用，另一方面，当所述支架处于折叠状态时也可以进一步缩小体积，增强用户体验。
以上根据图14A以及图14B对可用于所述视障辅助设备的支架例如桌面式支架的可能结构进行了示意说明。要注意的是，图14A以及图14B所示出的结构仅是一个示例，根据具体的实现方式，本公开的支架可以仅包括图14A或图14B所示出的组成部分中的一种或多个，或者还可包括其它更多的组件。
根据一些实施例，所述视障辅助设备还可包括被配置为对所述图像中包含的文字进行文字检测和识别以获得文字数据的电路；被配置为将所述文字数据转换成声音数据的电路；以及被配置为输出所述声音数据和/或所述文字数据的电路，等等。
根据本公开的另一方面，还提供一种电子设备，包括：处理器；以及存储程序的存储器，所述程序包括指令，所述指令在由所述处理器执行时使所述处理器执行上述的方法。
根据本公开的另一方面，还提供一种存储程序的非暂态计算机可读存储介质，所述程序包括指令，所述指令在由电子设备的处理器执行时，致使所述电子设备执行上述的方法。
图15是示出根据本公开的示例性实施例的电子设备的示例的框图。要注意的是，图15所示出的结构仅是一个示例，根据具体的实现方式，本公开的电子设备可以仅包括图15所示出的组成部分中的一种或多个。
电子设备2000例如可以是通用计算机(例如膝上型计算机、平板计算机等等各种计算机)、移动电话、个人数字助理。根据一些实施例，电子设备2000可以是视障辅助设备。
电子设备2000可被配置为采集或拍摄图像，并对所采集的(例如预览)图像进行处理，并且响应于所述处理而提供提示。例如，电子设备2000可被配置为采集图像，对该图像进行文字检测和识别以获得文字数据，将文字数据转换成声音数据，并且可以输出声音数据供用户聆听，和/或输出文字数据供用户观看。
根据一些实施方式，所述电子设备2000可以被配置为包括眼镜架或者被配置为能够可拆卸地安装到眼镜架(例如眼镜架的镜框、连接两个镜框的连接件、镜腿或任何其他部分)上，从而能够采集或拍摄到近似包括用户的视野的图像。
根据一些实施方式，所述电子设备2000也可被安装到其它可穿戴设备上，或者与其它可穿戴设备集成为一体。所述可穿戴设备例如可以是：头戴式设备(例如头盔或帽子等)、可佩戴在耳朵上的设备等。根据一些实施例，所述电子设备可被实施为可附接到可穿戴设备上的配件，例如可被实施为可附接到头盔或帽子上的配件等。
根据一些实施方式，所述电子设备2000也可具有其他形式。例如，电子设备2000可以是移动电话、通用计算设备(例如膝上型计算机、平板计算机等)、个人数字助理，等等。电子设备2000也可以具有底座，从而能够被安放在桌面上。
根据一些实施方式，所述电子设备2000作为视障辅助设备可以用于辅助阅读，在这种情况下，所述电子设备2000有时也被称为“电子阅读器”或“阅读辅助设备”。借助于电子设备2000，无法自主阅读的用户(例如视力障碍人士、存在阅读障碍的人士等)可以采用类似阅读姿势的姿势即可实现对常规读物(例如书本、杂志等)的“阅读”。在“阅读”过程中，所述电子设备2000可以获取采集的图像，并对所述图像进行检测，确定所述图像中是否存在包围所述读物的边界框以确认该读物是否被采集进所述图像，并且如果确定存在边界框，但是所采集的读物的大小或位置不太合适，则可以基于边界框在图像中的位置，提示用户移动所述读物，以拍摄得到大小、位置合适的读物，并且对拍摄的大小、位置合适的读物进行文字检测和文字识别(例如利用光学文字识别OCR方法)，以获得其中的文字数据，并可以将文字数据转换成声音数据，通过扬声器或耳机等声音输出设备输出所述声音数据供用户聆听。
通过在不提供预览图像或者无法预览图像的情况下，向用户提供移动目标对象的各种详尽的提示，能够帮助电子设备2000更快、更准确地获取到位置、大小合适的目标对象，从而能够更快、更准确地识别和输出给用户，能够大大提升响应速度、识别效率以及识别准确度。
电子设备2000可以包括图像传感器2004，用于采集、拍摄、获取图像。图像传感器2004可以采集和/或拍摄静态的图像，也可以采集和/或拍摄动态的图像，可以包括但不限于摄像头、照相机、视频摄像机等，被配置为获取包括目标对象的初始图像。电子设备2000还可以包括电子电路2100，所述电子电路2100包括被配置为执行如前所述的方法的步骤的电路。电子设备2100还可以包括文字识别电路2005，所述文字识别电路2005被配置为对所述图像中的文字进行文字检测和识别(例如OCR处理)，从而获得文字数据。所述文字识别电路2005例如可以通过专用芯片实现。电子设备2000还可以包括声音转换电路2006，所述声音转换电路2006被配置为将所述文字数据转换成声音数据。所述声音转换电路2006例如可以通过专用芯片实现。电子设备2000还可以包括声音输出电路2007，所述声音输出电路2007被配置为输出所述声音数据。所述声音输出电路2007可以包括但不限于耳机、扬声器、或振动器等，及其相应驱动电路。
根据一些实施方式，所述电子设备2000还可以包括图像处理电路2008，所述图像处理电路2008可以包括被配置为对图像进行各种图像处理的电路。图像处理电路2008例如可以包括但不限于以下中的一个或多个：被配置为对图像进行降噪的电路、被配置为对图像进行去模糊化的电路、被配置为对图像进行几何校正的电路、被配置为对图像进 行特征提取的电路、被配置为对图像中的目标对象进行目标检测和识别的电路、被配置为对图像中包含的文字进行文字检测的电路、被配置为从图像中提取文本行的电路、被配置为从图像中提取文字坐标的电路、被配置为从图像中提取边界框的电路、被配置为从图像中提取文本框的电路、被配置为基于图像进行版面分析(例如段落划分)的电路，等等。
根据一些实施方式，电子电路2100还可以包括文字处理电路2009，所述文字处理电路2009可以被配置为基于所提取的与文字有关的信息(例如文字数据、文本框、段落坐标、文本行坐标、文字坐标等)进行各种处理，从而获得诸如段落排序、文字语义分析、版面分析结果等处理结果。
上述的各种电路(例如文字识别电路2005、声音转换电路2006、声音输出电路2007、图像处理电路2008、文字处理电路2009、电子电路2100中的一个或多个可以使用定制硬件，和/或可以用硬件、软件、固件、中间件、微代码，硬件描述语言或其任何组合来实现。例如，上述的各种电路中的一个或多个可以通过使用根据本公开的逻辑和算法，用汇编语言或硬件编程语言(诸如VERILOG，VHDL，C++)对硬件(例如，包括现场可编程门阵列(FPGA)和/或可编程逻辑阵列(PLA)的可编程逻辑电路)进行编程来实现。
根据一些实施方式，电子设备2000还可以包括通信电路2010，所述通信电路2010可以是使得能够与外部设备和/或与网络通信的任何类型的设备或系统，并且可以包括但不限于调制解调器、网卡、红外通信设备、无线通信设备和/或芯片组，例如蓝牙设备、1302.11设备、WiFi设备、WiMax设备、蜂窝通信设备和/或类似物。
根据一些实施方式，电子设备2000还可以包括输入设备2011，所述输入设备2011可以是能向电子设备2000输入信息的任何类型的设备，并且可以包括但不限于各种传感器、鼠标、键盘、触摸屏、按钮、控制杆、麦克风和/或遥控器等等。
根据一些实施方式，电子设备2000还可以包括输出设备2012，所述输出设备2012可以是能呈现信息的任何类型的设备，并且可以包括但不限于显示器、视觉输出终端、振动器和/或打印机等。尽管电子设备2000根据一些实施例用于视障辅助设备，基于视觉的输出设备可以方便用户的家人或维修工作人员等从电子设备2000获得输出信息。
根据一些实施方式，电子设备2000还可以包括处理器2001。所述处理器2001可以是任何类型的处理器，并且可以包括但不限于一个或多个通用处理器和/或一个或多个专用处理器(例如特殊处理芯片)。处理器2001例如可以是但不限于中央处理单元CPU 或微处理器MPU等等。电子设备2000还可以包括工作存储器2002，所述工作存储器2002可以存储对处理器2001的工作有用的程序(包括指令)和/或数据(例如图像、文字、声音，以及其他中间数据等)的工作存储器，并且可以包括但不限于随机存取存储器和/或只读存储器设备。电子设备2000还可以包括存储设备2003，所述存储设备2003可以包括任何非暂时性存储设备，非暂时性存储设备可以是非暂时性的并且可以实现数据存储的任何存储设备，并且可以包括但不限于磁盘驱动器、光学存储设备、固态存储器、软盘、柔性盘、硬盘、磁带或任何其他磁介质，光盘或任何其他光学介质、ROM(只读存储器)、RAM(随机存取存储器)、高速缓冲存储器和/或任何其他存储器芯片或盒、和/或计算机可从其读取数据、指令和/或代码的任何其他介质。工作存储器2002和存储设备2003可以被集合地称为“存储器”，并且在有些情况下可以相互兼用。
根据一些实施方式，处理器2001可以对图像传感器2004、文字识别电路2005、声音转换电路2006、声音输出电路2007、图像处理电路2008、文字处理电路2009、通信电路2010、电子电路2100以及电子设备2000包括的其他各种装置和电路中的至少一个进行控制和调度。根据一些实施方式，图15中所述的各个组成部分中的至少一些可通过线路2013而相互连接和/或通信。
软件要素(程序)可以位于所述工作存储器2002中，包括但不限于操作系统2002a、一个或多个应用程序2002b、驱动程序和/或其他数据和代码。
根据一些实施方式，用于进行前述的控制和调度的指令可以被包括在操作系统2002a或者一个或多个应用程序2002b中。
根据一些实施方式，执行本公开所述的方法步骤的指令可以被包括在一个或多个应用程序2002b中，并且上述电子设备2000的各个模块可以通过由处理器2001读取和执行一个或多个应用程序2002b的指令来实现。换言之，电子设备2000可以包括处理器2001以及存储程序的存储器(例如工作存储器2002和/或存储设备2003)，所述程序包括指令，所述指令在由所述处理器2001执行时使所述处理器2001执行如本公开各种实施例所述的方法。
根据一些实施方式，文字识别电路2005、声音转换电路2006、图像处理电路2008、文字处理电路2009、电子电路2100中的至少一个所执行的操作中的一部分或者全部可以由处理器2001读取和执行一个或多个应用程序2002的指令来实现。
软件要素(程序)的指令的可执行代码或源代码可以存储在非暂时性计算机可读存储介质(例如所述存储设备2003)中，并且在执行时可以被存入工作存储器2001中(可 能被编译和/或安装)。因此，本公开提供存储程序的计算机可读存储介质，所述程序包括指令，所述指令在由电子设备(例如视障辅助设备)的处理器执行时，致使所述电子设备执行如本公开各种实施例所述的方法。根据另一种实施方式，软件要素(程序)的指令的可执行代码或源代码也可以从远程位置下载。
还应该理解，可以根据具体要求而进行各种变型。例如，也可以使用定制硬件，和/或可以用硬件、软件、固件、中间件、微代码，硬件描述语言或其任何组合来实现各个电路、单元、模块或者元件。例如，所公开的方法和设备所包含的电路、单元、模块或者元件中的一些或全部可以通过使用根据本公开的逻辑和算法，用汇编语言或硬件编程语言(诸如VERILOG，VHDL，C++)对硬件(例如，包括现场可编程门阵列(FPGA)和/或可编程逻辑阵列(PLA)的可编程逻辑电路)进行编程来实现。
根据一些实施方式，电子设备2000中的处理器2001可以分布在网络上。例如，可以使用一个处理器执行一些处理，而同时可以由远离该一个处理器的另一个处理器执行其他处理。电子设备2001的其他模块也可以类似地分布。这样，电子设备2001可以被解释为在多个位置执行处理的分布式计算系统。
虽然已经参照附图描述了本公开的实施例或示例，但应理解，上述的方法、系统和设备仅仅是示例性的实施例或示例，本公开的范围并不由这些实施例或示例限制，而是仅由授权后的权利要求书及其等同范围来限定。实施例或示例中的各种要素可以被省略或者可由其等同要素替代。此外，可以通过不同于本公开中描述的次序来执行各步骤。进一步地，可以用各种方式组合实施例或示例中的各种要素。重要的是随着技术的演进，在此描述的很多要素可以由本公开之后出现的等同要素进行替换。

Claims (21)

  1. 一种计算机实现的目标检测方法，包括：
    获取图像传感器采集的图像；
    检测所述图像中是否存在目标对象的至少一部分；以及
    响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。
  2. 如权利要求1所述的目标检测方法，还包括：
    在提供调整所述目标对象至所述图像传感器视场范围内的提示之后，重新获取图像并继续检测重新获取的图像中是否存在所述目标对象的至少一部分，直到在重新获取的图像中检测到所述目标对象的至少一部分。
  3. 如权利要求1所述的目标检测方法，其中，所述提供调整所述目标对象至所述图像传感器视场范围内的提示，包括：
    响应于所述图像传感器为位置固定的设备，提供移动所述目标对象以使所述目标对象进入所述图像传感器视场范围内的提示；以及
    响应于所述图像传感器为位置可调的设备，提供移动所述目标对象和/或所述图像传感器以使所述目标对象进入所述图像传感器视场范围内的提示。
  4. 如权利要求1～3中的任一项所述的目标检测方法，还包括：
    响应于在所述图像中检测到所述目标对象的至少一部分，确定所述目标对象的所述至少一部分与所述图像的几何关系是否满足特定条件；以及
    响应于所述目标对象的所述至少一部分与所述图像的几何关系满足所述特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
  5. 如权利要求4所述的目标检测方法，其中，所述目标对象的所述至少一部分由在所述图像中检测到的对应的边界框来表示，以及其中，所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    响应于所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
  6. 如权利要求5所述的目标检测方法，其中，所述边界框为矩形框，所述特定距离要求包括所述矩形框的每条边与所述图像的相应边之间的距离均大于相应的特定距离阈值，以及
    其中，所述响应于所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    响应于检测到的所述矩形框的每条边与所述图像的相应边之间的距离均小于相应的特定距离阈值，直接提示使目标对象远离；以及
    响应于检测到的所述矩形框的边与所述图像的相应边之间的四个距离之中有三个或者更少数量的距离小于相应的特定距离阈值，提示使目标对象远离，和/或将所述目标对象沿下述方位移动：该方位是沿所述边界框的中心位置相对于所述图像的中心位置或中心区域的方位相反的方位。
  7. 如权利要求5所述的目标检测方法，其中，所述边界框为矩形框，所述特定距离要求还包括所述边界框的边与所述图像的相应边之间的距离在相应的特定距离范围内，以及其中，所述响应于所述边界框的边与所述图像的相应边之间的距离不符合特定距离要求，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    响应于所述边界框的边与所述图像的相应边之间的距离中至少一个小于相应的特定距离范围的下限，提示使目标对象远离和/或沿所述边界框的小于所述相应的特定距离范围下限的边所在方位相反的方向使目标对象偏移；以及
    响应于所述边界框的边与所述图像的相应边之间的距离中至少一个大于相应的特定距离范围的上限，提示使目标对象靠近和/或沿与所述边界框的大于所述相应的特定距离范围的上限的边所在方位一致的方向使目标对象偏移。
  8. 如权利要求4所述的目标检测方法，其中，所述目标对象的所述至少一部分由在所述图像中检测到的对应的边界框来表示，以及其中，所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    响应于所述边界框的中心位置不在所述图像的中心区域中，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
  9. 如权利要求8所述的目标检测方法，其中，所述响应于所述边界框的中心位置不在所述图像的中心区域中，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    基于所述边界框的中心位置相对于所述图像的中心区域的方位，提示沿与该方位相反的方向使所述目标对象移动。
  10. 如权利要求4所述的目标检测方法，其中，所述目标对象的所述至少一部分由在所述图像中检测到的对应的边界框来表示，以及其中，所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器的相对位置的提示，包括：
    响应于所述边界框的面积与所述图像的面积的比值不在特定比值范围内，提供调整所述目标对象与所述图像传感器之间的相对位置的提示。
  11. 如权利要求10所述的目标检测方法，其中，所述响应于所述边界框的面积与所述图像的面积的比值不在特定比值范围内，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    响应于所述边界框的面积与所述图像的面积的比值高于所述特定比值范围的上限，提示使目标对象放远离；
    响应于所述边界框的面积与所述图像的面积的比值低于所述特定比值范围的下限，提示使将目标对象靠近。
  12. 如权利要求4所述的目标检测方法，其中，所述目标对象中包含文本区域，
    所述响应于所述目标对象的所述至少一部分与所述图像的几何关系满足特定条件，提供调整所述目标对象与所述图像传感器之间的相对位置的提示，包括：
    确定所述目标对象的文本区域的文本行相对于所述图像的一条边的角度是否位于特定角度范围内，响应于该角度不位于所述特定角度范围内，提示调整所述目标对象与所述图像传感器之间的相对角度。
  13. 如权利要求2所述的目标检测方法，其中，所述重新获取的图像是图像传感器重新采集的图像，以及其中，使得图像传感器在至少间隔特定时间之后才重新采集图像。
  14. 如权利要求1所述的目标检测方法，其中，
    所获取的图像为通过所述图像传感器采集的预览图像，
    其中，所述预览图像的分辨率小于通过所述图像传感器采集的拍照图像的分辨率。
  15. 一种电子电路，包括：
    被配置为执行如权利要求1～14中任一项所述的方法的步骤的电路。
  16. 一种视障辅助设备，包括：
    图像传感器，被配置为采集图像；以及
    电子电路，被配置为检测所述图像中是否存在目标对象的至少一部分；其中，
    所述电子电路还被配置为响应于在所述图像中未检测到所述目标对象的至少一部分，提供调整所述目标对象至所述图像传感器视场范围内的提示。
  17. 如权利要求16所述的视障辅助设备，其中，
    所述电子电路被配置为响应于所述图像传感器为位置固定的设备，提供移动所述目标对象以使所述目标对象进入所述图像传感器视场范围内的提示；以及
    响应于所述图像传感器为位置可调的设备，提供移动所述目标对象和/或所述图像传感器以使所述目标对象进入所述图像传感器视场范围内的提示。
  18. 如权利要求16或17所述的视障辅助设备，其中，所述视障辅助设备包括下述设备中的一种或多种：
    可连接至可穿戴设备的视障辅助设备、手持式视障辅助设备以及桌面式视障辅助设备。
  19. 如权利要求16或17所述的视障辅助设备，其中，所述视障辅助设备还包括支架，所述支架包括连接件、伸缩立臂、以及用于支撑所述连接件以及所述伸缩立臂的带配重底座，其中，
    所述连接件，被配置为安装所述图像传感器以及所述电子电路；
    所述伸缩立臂的一端与所述连接件相连，另一端通过底座旋转关节与所述带配重底座相连，被配置为支撑所述连接件并通过伸缩和/旋转调节所述连接件相对于所述带配重底座的位置。
  20. 一种电子设备，包括：
    处理器；以及
    存储程序的存储器，所述程序包括指令，所述指令在由所述处理器执行时使所述处理器执行如权利要求1～14中任一项所述的方法。
  21. 一种存储程序的非暂态计算机可读存储介质，所述程序包括指令，所述指令在由电子设备的处理器执行时，致使所述电子设备执行根据权利要求1～14中任一项所述的方法。
PCT/CN2020/076152 2019-12-25 2020-02-21 目标检测方法、电路、视障辅助设备、电子设备和介质 WO2021128553A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/834,957 US10838056B1 (en) 2019-12-25 2020-03-30 Detection of target

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911353328.0 2019-12-25
CN201911353328.0A CN111163261B (zh) 2019-12-25 2019-12-25 目标检测方法、电路、视障辅助设备、电子设备和介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/834,957 Continuation US10838056B1 (en) 2019-12-25 2020-03-30 Detection of target

Publications (1)

Publication Number Publication Date
WO2021128553A1 true WO2021128553A1 (zh) 2021-07-01

Family

ID=70189656

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/076152 WO2021128553A1 (zh) 2019-12-25 2020-02-21 目标检测方法、电路、视障辅助设备、电子设备和介质

Country Status (4)

Country Link
EP (1) EP3842992B1 (zh)
JP (1) JP6800501B1 (zh)
CN (1) CN111163261B (zh)
WO (1) WO2021128553A1 (zh)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114071005B (zh) * 2020-08-07 2022-12-27 华为技术有限公司 一种对象检测方法、电子设备和计算机可读存储介质
CN112422829B (zh) * 2020-11-19 2022-04-26 北京字节跳动网络技术有限公司 辅助拍摄图像的方法、装置、终端和存储介质
WO2022222047A1 (zh) * 2021-04-20 2022-10-27 Oppo广东移动通信有限公司 文档扫描方法及装置、存储介质及电子设备
CN113411477B (zh) * 2021-06-10 2023-03-10 支付宝(杭州)信息技术有限公司 一种图像的采集方法、装置及设备
CN113255632B (zh) * 2021-07-16 2023-01-03 深圳市赛菲姆科技有限公司 基于车牌识别的摄像头参数调整方法、装置、设备和介质
CN114895832B (zh) * 2022-05-17 2023-08-08 网易（杭州）网络有限公司 对象的调整方法、装置、电子设备及计算机可读介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120281874A1 (en) * 2011-05-05 2012-11-08 Lure Yuan-Ming F Method, material, and apparatus to improve acquisition of human frontal face images using image template
CN103813075A (zh) * 2012-11-07 2014-05-21 联想(北京)有限公司 一种提醒方法和电子设备
CN104253938A (zh) * 2013-06-26 2014-12-31 中兴通讯股份有限公司 终端及其智能拍照的方法
CN107404721A (zh) * 2016-05-20 2017-11-28 阿里巴巴集团控股有限公司 物联网设备配网方法、图像采集方法及设备

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102096927A (zh) * 2011-01-26 2011-06-15 北京林业大学 自主林业机器人目标跟踪方法
CN103139480A (zh) * 2013-02-28 2013-06-05 华为终端有限公司 一种图像采集方法及装置
JP6700661B2 (ja) * 2015-01-30 2020-05-27 キヤノン株式会社 画像処理装置、画像処理方法、及び画像処理システム
CN107645628B (zh) * 2016-07-21 2021-08-06 中兴通讯股份有限公司 一种信息处理方法及装置
US10382673B2 (en) * 2017-01-11 2019-08-13 Hover Inc. Automated guide for image capturing for 3D model creation

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120281874A1 (en) * 2011-05-05 2012-11-08 Lure Yuan-Ming F Method, material, and apparatus to improve acquisition of human frontal face images using image template
CN103813075A (zh) * 2012-11-07 2014-05-21 联想(北京)有限公司 一种提醒方法和电子设备
CN104253938A (zh) * 2013-06-26 2014-12-31 中兴通讯股份有限公司 终端及其智能拍照的方法
CN107404721A (zh) * 2016-05-20 2017-11-28 阿里巴巴集团控股有限公司 物联网设备配网方法、图像采集方法及设备

Also Published As

Publication number Publication date
CN111163261A (zh) 2020-05-15
JP6800501B1 (ja) 2020-12-16
EP3842992A1 (en) 2021-06-30
CN111163261B (zh) 2022-03-01
JP2021103503A (ja) 2021-07-15
EP3842992B1 (en) 2023-06-21

Similar Documents

Publication Publication Date Title
WO2021128553A1 (zh) 目标检测方法、电路、视障辅助设备、电子设备和介质
US9712751B2 (en) Camera field of view effects based on device orientation and scene content
EP3968625B1 (en) Digital photographing apparatus and method of operating the same
KR102429427B1 (ko) 촬영 장치 및 그 동작 방법
EP3188467A1 (en) Method for image capturing using unmanned image capturing device and electronic device supporting the same
US10838056B1 (en) Detection of target
White et al. EasySnap: real-time audio feedback for blind photography
US10291843B2 (en) Information processing apparatus having camera function and producing guide display to capture character recognizable image, control method thereof, and storage medium
US9922250B2 (en) Method of capturing iris image, computer-readable recording medium storing the method, and iris image capturing apparatus
US20140185957A1 (en) Apparatus and method for processing image in device having camera
KR102402148B1 (ko) 전자 장치 및 그의 문자 인식 방법
JP6096654B2 (ja) 画像の記録方法、電子機器およびコンピュータ・プログラム
CN111126394A (zh) 文字识别方法、阅读辅助设备、电路和介质
US20110007191A1 (en) Apparatus and method for processing digital image
CN114208150A (zh) 用于提供相机预览图像的电子装置及其操作方法
WO2022121842A1 (zh) 文本图像的矫正方法及装置、设备和介质
WO2018196854A1 (zh) 一种拍照方法、拍照装置及移动终端
CN110869878A (zh) 用户信号处理方法及执行该方法的装置
EP4184931A1 (en) Image capturing apparatus capable of suppressing detection of subject not intended by user, control method for image capturing apparatus, and storage medium
WO2022121843A1 (zh) 文本图像的矫正方法及装置、设备和介质
CN111145153A (zh) 图像处理方法、电路、视障辅助设备、电子设备及介质
US10902265B2 (en) Imaging effect based on object depth information
JP6450604B2 (ja) 画像取得装置及び画像取得方法
US11086194B2 (en) Camera accessory mask
JP2011113196A (ja) 顔方向特定装置及び撮像装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20908242

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20908242

Country of ref document: EP

Kind code of ref document: A1