CN111191730A - Method and system for detecting oversized image target facing embedded deep learning - Google Patents
Method and system for detecting oversized image target facing embedded deep learning Download PDFInfo
- Publication number
- CN111191730A CN111191730A CN202010003131.0A CN202010003131A CN111191730A CN 111191730 A CN111191730 A CN 111191730A CN 202010003131 A CN202010003131 A CN 202010003131A CN 111191730 A CN111191730 A CN 111191730A
- Authority
- CN
- China
- Prior art keywords
- image
- target
- sub
- target detection
- deep learning
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides an embedded deep learning-oriented ultra-large image target detection method and system, which comprises the following steps: the device comprises an image preprocessing unit, a target detection unit and an image post-processing unit. The method for detecting the oversized image target facing the embedded deep learning, provided by the invention, faces to more and more oversized image processing requirements in the field of the embedded deep learning, and aims at the limitation of an embedded multi-core processor running deep neural network, a single image is divided into a plurality of images based on a blocking idea, so that the parallel target detection of the single image is realized, the detection result is analyzed by an image post-processing unit, the plurality of detection results are integrated, and the problem of low processing efficiency of an embedded computing platform is effectively solved.
Description
Technical Field
The invention belongs to the field of intelligent calculation, and relates to an ultra-large image target detection method for embedded deep learning.
Background
With the continuous development of deep learning technology, image processing algorithms based on deep learning, such as target detection, semantic segmentation and the like, have achieved great success. However, due to the complex structure and numerous parameters of the deep neural network, the deep learning algorithm is poor in performance on an embedded computing platform with limited resources, and can only process a few small images. For large-size remote sensing images or high-definition images rich in detail information, the processing efficiency of the embedded deep learning algorithm is very low.
In order to improve the processing efficiency of the embedded deep learning algorithm, hardware manufacturers successively release a plurality of AI chips, the processing speed is improved by optimizing a computing structure on a hardware level, and the data throughput is improved by a multi-core technology. However, due to the high interconnectivity of the deep neural network, the parallelization computation difficulty of the algorithm model for processing a single image is increased. The real-time processing requirement for streaming media data is difficult to fully utilize the parallel computing capability of a multi-core processor.
Disclosure of Invention
The invention provides an embedded deep learning-oriented ultra-large image target detection method, which aims to efficiently process ultra-large image data in real time and complete comprehensive and accurate detection of a target.
The solution proposed by the invention is as follows:
the method for detecting the oversized image target facing the embedded deep learning comprises the following steps:
1) receiving an input image, and dividing the image into a plurality of sub-images according to the positions of pixel points, wherein any one sub-image is mutually overlapped with all the adjacent sub-images in a region close to a boundary, and the region is marked as a division redundant region;
2) respectively carrying out target detection on each sub-image to obtain target related information (target detection result);
3) the obtained target related information of each sub-image is referred to, and the targets of the divided redundant areas are re-detected and positioned; and marking the original image according to the updated target detection result, and outputting a visualization result.
Based on the above scheme, the invention further optimizes as follows:
optionally, in the step 1), image segmentation is specifically performed according to a preset width W, a preset height H, and a preset redundancy threshold T; the redundancy threshold value T characterizes the number of pixels which are overlapped with each other in the area close to the boundary; and 2) specifically, respectively carrying out target detection on the sub-images with the size of W multiplied by H based on a convolutional neural network algorithm, and outputting target related information, wherein the target related information at least comprises a target position.
Optionally, the target related information further includes an affiliated category and a confidence level.
Optionally, the width W, the height H and the redundancy threshold T are determined according to the original image size, the target size and the processor computing power;
the width range of the divided image is [0, W-1], [ W-T,2W-T-1], [2W-2T,3W-2T-1] … …;
the height range of the segmentation image is [0, H-1], [ H-T,2H-T-1], [2H-2T,3H-2T-1] … …;
any width range and height range together constitute a sub-image area.
Optionally, step 3) specifically includes:
analyzing whether a target regression frame exists in the segmentation redundant area of each sub-image, wherein the distance from the target regression frame to the boundary of the sub-image is less than aT, a is a set coefficient, and the preferable value range is 0< a < 0.5;
if the target regression frame meeting the conditions exists, re-determining the corresponding width or height range by taking the segmentation redundant area as the center, and sampling a new sub-image with the size of W multiplied by H;
and carrying out target detection on the new sub-image again, only adopting a target detection result of the segmented redundant area, updating the target related information of the segmented redundant area, and marking all the target related information on the original image according to a required rule according to the target related information of the previous non-segmented redundant area to form a visual output image.
Correspondingly, the invention also provides an embedded deep learning-oriented oversized image target detection system, which comprises:
the image preprocessing unit is used for receiving an input image and dividing the image into a plurality of sub-images according to the positions of pixel points, wherein any one sub-image is mutually overlapped with all the adjacent sub-images in a region close to a boundary, and the region is marked as a division redundant region;
the target detection unit is used for respectively carrying out target detection on each sub-image to obtain target related information;
the image post-processing unit is used for detecting and positioning the targets of the divided redundant areas again by referring to the obtained target related information of each sub-image; and marking the original image according to the updated target detection result, and outputting a visualization result.
Optionally, the image post-processing unit detects and positions the target of the segmented redundant area again, specifically: and generating a new sub-image by taking the divided redundant area as a center, reentering the target detection unit for target detection, and only adopting a target detection result of the divided redundant area.
Correspondingly, the invention also provides an embedded device, which comprises a processor and a program memory, wherein when a program stored in the program memory is loaded by the processor, the method for detecting the ultra-large-size image target facing the embedded deep learning is executed.
Compared with the prior art, the invention has the following beneficial effects:
the method for detecting the target of the super-large-size image facing the embedded deep learning realizes the parallel target detection of a single super-large-size image based on the deep neural network, fully utilizes the computing resources of a multi-core processor, improves the deep learning processing efficiency under the embedded computing environment, and effectively solves the real-time processing problem of the captured data of the (super) high-definition camera.
The invention fully considers that the possible target is on the segmentation line when the sub-image is segmented, so that the redundant processing (which is equivalent to the offset of the standard grid unit in the horizontal direction and the vertical direction) and the corresponding image post-processing are particularly carried out on the pixels near the boundary, thereby avoiding the missing detection of the target and realizing the complete and accurate detection of the target.
Drawings
Fig. 1 is a schematic diagram of the principle of the present invention.
Detailed Description
The invention is further described in detail below with reference to the figures and examples.
The invention provides an embedded depth learning-oriented oversized image target detection method aiming at the increasing requirements of the field of embedded depth learning and aiming at the limitation of an embedded multi-core processor running depth neural network.
As shown in fig. 1, the method for detecting an image target with a super-large size facing embedded deep learning may be implemented by the following software modules:
an image pre-processing unit to: receiving an input image, and performing image segmentation according to a preset width W, a preset height H and a preset redundancy threshold value T. Specifically, the image width is divided into [0, W-1], [ W-T,2W-T-1], [2W-2T,3W-2T-1] … …, and the image height is divided into [0, H-1], [ H-T,2H-T-1], [2H-2T,3H-2T-1] … …;
the target detection unit is used for: respectively carrying out target detection on the sub-images with the size of WxH based on a convolutional neural network algorithm, and marking information such as target positions, categories and confidence degrees;
the image post-processing unit is used for: and fusing target detection results, and re-detecting and positioning the targets of the divided redundant areas to form a visual target detection result. Specifically, whether a target regression frame exists in the adjacent sub-image segmentation redundant area or not is analyzed, and the distance from the target regression frame to the sub-image boundary is smaller than aT (0< a <0.5), and then target detection is carried out again on the sub-image with the segmentation redundant area as the center and the sampling size of W multiplied by H; and taking the secondary detection result as the standard for the target in the partitioned redundant area.
The image post-processing unit can update the relevant information of the target of the segmented redundant area according to the second round of target detection results, and mark the relevant information of the target of the non-segmented redundant area on the original image according to the required rule according to the relevant information of the target of the non-segmented redundant area in the first round to form a visual output image.
Taking the Faster R-CNN target detection algorithm as an example, the specific working flow is shown in FIG. 1, and the specific process is as follows:
first, an input image is segmented by an image preprocessing unit based on a preset rule to obtain a group of sub-images. The widths and heights of the sub-images are set to 612 and 426, respectively, and the redundancy threshold is 30, that is, the width of the original image is divided into [0,611] [582,1193] [1164,1775] [1746,2357] [2328,2939] [2910,3521] [3492,4103], [4096,4103] is filled with zero expansion, the height of the original image is divided into [0,455] [426,881] [852,1307] [1278,1733] [1704,2159], and the total of 7 × 5 to 35 sub-images, by comprehensively considering the original image size (4096 × 2160), the target size (assuming that the minimum target size is 32 × 32 and the maximum target size is 200 × 200), the computing power of the processor, and the like.
Then, the subimages enter a target detection unit in batch, and the target detection is carried out in parallel based on the Faster R-CNN algorithm.
Finally, the image post-processing unit analyzes the target detection result, and judges whether a target regression frame is in the segmentation redundant area and the distance between the target regression frame and the edge of the sub-image is less than 9 pixels (0.3T); if the target meeting the conditions exists, sampling and generating a new sub-image by taking the divided redundant area as the center, and detecting the target again. For example, if the target regression frame diagonal of the sub-image 2 is ([585,215] [645,255]), that is, the regression frame is 3 pixel distances away from the edge of the sub-image, the area of the abscissa range [241,852] and the ordinate range [0,455] is resampled to generate a new sub-image, and the new sub-image enters the target detection unit to reposition the target position in the middle area. In addition, the image post-processing unit marks target information such as positions, the categories, confidence degrees and the like on the original image according to the target detection result, completes visualization processing and outputs the image.
For the above target detection of each sub-image, different conventional algorithms (usually based on convolutional neural network algorithm) may be adopted according to actual needs, for example, besides the fast R-CNN algorithm (focusing on positioning accuracy), the YOLO algorithm (focusing on operation speed) may also be adopted.
Claims (9)
1. An embedded deep learning-oriented ultra-large image target detection method is characterized by comprising the following steps:
1) receiving an input image, and dividing the image into a plurality of sub-images according to the positions of pixel points, wherein any one sub-image is mutually overlapped with all the adjacent sub-images in a region close to a boundary, and the region is marked as a division redundant region;
2) respectively carrying out target detection on each sub-image to obtain target related information;
3) the obtained target related information of each sub-image is referred to, and the targets of the divided redundant areas are re-detected and positioned; and marking the original image according to the updated target detection result, and outputting a visualization result.
2. The embedded deep learning-oriented oversized image target detection method according to claim 1, characterized by:
step 1) specifically, image segmentation is carried out according to a preset width W, a preset height H and a preset redundancy threshold value T; the redundancy threshold value T characterizes the number of pixels which are overlapped with each other in the area close to the boundary;
and 2) specifically, respectively carrying out target detection on the sub-images with the size of W multiplied by H based on a convolutional neural network algorithm, and outputting target related information, wherein the target related information at least comprises a target position.
3. The embedded deep learning-oriented oversized image target detection method according to claim 2, characterized by comprising the following steps: the target related information further comprises a category to which the target related information belongs and a confidence level.
4. The embedded deep learning-oriented oversized image target detection method according to claim 2, characterized by comprising the following steps: the width W, the height H and the redundancy threshold value T are determined according to the original image size, the target size and the calculation capacity of a processor;
the width range of the divided image is [0, W-1], [ W-T,2W-T-1], [2W-2T,3W-2T-1] … …;
the height range of the segmentation image is [0, H-1], [ H-T,2H-T-1], [2H-2T,3H-2T-1] … …;
any width range and height range together constitute a sub-image area.
5. The embedded deep learning-oriented oversized image target detection method according to claim 2, characterized by comprising the following steps: the step 3) specifically comprises the following steps:
analyzing whether a target regression frame exists in the segmentation redundant area of each sub-image, wherein the distance from the target regression frame to the boundary of the sub-image is less than aT, and a is a set coefficient;
if the target regression frame meeting the conditions exists, re-determining the corresponding width or height range by taking the segmentation redundant area as the center, and sampling a new sub-image with the size of W multiplied by H;
and carrying out target detection on the new sub-image again, only adopting a target detection result of the segmented redundant area, updating the target related information of the segmented redundant area, and marking all the target related information on the original image according to a required rule according to the target related information of the previous non-segmented redundant area to form a visual output image.
6. The embedded deep learning-oriented oversized image target detection method according to claim 5, characterized by comprising the following steps: 0< a < 0.5.
7. An embedded deep learning oriented ultra-large image target detection system is characterized by comprising:
the image preprocessing unit is used for receiving an input image and dividing the image into a plurality of sub-images according to the positions of pixel points, wherein any one sub-image is mutually overlapped with all the adjacent sub-images in a region close to a boundary, and the region is marked as a division redundant region;
the target detection unit is used for respectively carrying out target detection on each sub-image to obtain target related information;
the image post-processing unit is used for detecting and positioning the targets of the divided redundant areas again by referring to the obtained target related information of each sub-image; and marking the original image according to the updated target detection result, and outputting a visualization result.
8. The system for detecting the targets of the oversized images facing the embedded deep learning as claimed in claim 7, wherein the image post-processing unit is used for detecting and positioning the targets of the segmented redundant regions again, and specifically comprises: and generating a new sub-image by taking the divided redundant area as a center, reentering the target detection unit for target detection, and only adopting a target detection result of the divided redundant area.
9. An embedded device comprising a processor and a program memory, wherein the program stored in the program memory is loaded by the processor to execute the method for detecting the object of the ultra-large size image facing the embedded deep learning of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010003131.0A CN111191730B (en) | 2020-01-02 | 2020-01-02 | Method and system for detecting oversized image target oriented to embedded deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010003131.0A CN111191730B (en) | 2020-01-02 | 2020-01-02 | Method and system for detecting oversized image target oriented to embedded deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111191730A true CN111191730A (en) | 2020-05-22 |
| CN111191730B CN111191730B (en) | 2023-05-12 |
Family
ID=70709746
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010003131.0A Active CN111191730B (en) | 2020-01-02 | 2020-01-02 | Method and system for detecting oversized image target oriented to embedded deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111191730B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132164A (en) * | 2020-11-20 | 2020-12-25 | 北京易真学思教育科技有限公司 | Target detection method, system, computer device and storage medium |
| CN113762220A (en) * | 2021-11-03 | 2021-12-07 | 通号通信信息集团有限公司 | Object recognition method, electronic device, and computer-readable storage medium |
| CN114332456A (en) * | 2022-03-16 | 2022-04-12 | 山东力聚机器人科技股份有限公司 | Target detection and identification method and device for large-resolution image |
| WO2023116641A1 (en) * | 2021-12-21 | 2023-06-29 | 北京罗克维尔斯科技有限公司 | Parking space detection model training method and apparatus, and parking space detection method and apparatus |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6751363B1 (en) * | 1999-08-10 | 2004-06-15 | Lucent Technologies Inc. | Methods of imaging based on wavelet retrieval of scenes |
| US20150036921A1 (en) * | 2013-08-02 | 2015-02-05 | Canon Kabushiki Kaisha | Image composition evaluating apparatus, information processing apparatus and methods thereof |
| CN104408482A (en) * | 2014-12-08 | 2015-03-11 | 电子科技大学 | Detecting method for high-resolution SAR (Synthetic Aperture Radar) image object |
| JP2015106360A (en) * | 2013-12-02 | 2015-06-08 | 三星電子株式会社Ｓａｍｓｕｎｇ Ｅｌｅｃｔｒｏｎｉｃｓ Ｃｏ．，Ｌｔｄ． | Object detection method and object detection device |
| US20160012302A1 (en) * | 2013-03-21 | 2016-01-14 | Fuji Xerox Co., Ltd. | Image processing apparatus, image processing method and non-transitory computer readable medium |
| US20160328856A1 (en) * | 2015-05-08 | 2016-11-10 | Qualcomm Incorporated | Systems and methods for reducing a plurality of bounding regions |
| US20170357860A1 (en) * | 2016-06-09 | 2017-12-14 | Hyundai Motor Company | Method and apparatus for detecting side of object using ground boundary information of obstacle |
| WO2018058573A1 (en) * | 2016-09-30 | 2018-04-05 | 富士通株式会社 | Object detection method, object detection apparatus and electronic device |
| CN108154521A (en) * | 2017-12-07 | 2018-06-12 | 中国航空工业集团公司洛阳电光设备研究所 | A kind of moving target detecting method based on object block fusion |
| KR101896357B1 (en) * | 2018-02-08 | 2018-09-07 | 주식회사 라디코 | Method, device and program for detecting an object |
| KR20180107988A (en) * | 2017-03-23 | 2018-10-04 | 한국전자통신연구원 | Apparatus and methdo for detecting object of image |
| WO2019000653A1 (en) * | 2017-06-30 | 2019-01-03 | 清华大学深圳研究生院 | Image target identification method and apparatus |
| CN110781839A (en) * | 2019-10-29 | 2020-02-11 | 北京环境特性研究所 | Sliding window-based small and medium target identification method in large-size image |
-
2020
- 2020-01-02 CN CN202010003131.0A patent/CN111191730B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6751363B1 (en) * | 1999-08-10 | 2004-06-15 | Lucent Technologies Inc. | Methods of imaging based on wavelet retrieval of scenes |
| US20160012302A1 (en) * | 2013-03-21 | 2016-01-14 | Fuji Xerox Co., Ltd. | Image processing apparatus, image processing method and non-transitory computer readable medium |
| US20150036921A1 (en) * | 2013-08-02 | 2015-02-05 | Canon Kabushiki Kaisha | Image composition evaluating apparatus, information processing apparatus and methods thereof |
| JP2015106360A (en) * | 2013-12-02 | 2015-06-08 | 三星電子株式会社Ｓａｍｓｕｎｇ Ｅｌｅｃｔｒｏｎｉｃｓ Ｃｏ．，Ｌｔｄ． | Object detection method and object detection device |
| CN104408482A (en) * | 2014-12-08 | 2015-03-11 | 电子科技大学 | Detecting method for high-resolution SAR (Synthetic Aperture Radar) image object |
| US20160328856A1 (en) * | 2015-05-08 | 2016-11-10 | Qualcomm Incorporated | Systems and methods for reducing a plurality of bounding regions |
| US20170357860A1 (en) * | 2016-06-09 | 2017-12-14 | Hyundai Motor Company | Method and apparatus for detecting side of object using ground boundary information of obstacle |
| WO2018058573A1 (en) * | 2016-09-30 | 2018-04-05 | 富士通株式会社 | Object detection method, object detection apparatus and electronic device |
| KR20180107988A (en) * | 2017-03-23 | 2018-10-04 | 한국전자통신연구원 | Apparatus and methdo for detecting object of image |
| WO2019000653A1 (en) * | 2017-06-30 | 2019-01-03 | 清华大学深圳研究生院 | Image target identification method and apparatus |
| CN108154521A (en) * | 2017-12-07 | 2018-06-12 | 中国航空工业集团公司洛阳电光设备研究所 | A kind of moving target detecting method based on object block fusion |
| KR101896357B1 (en) * | 2018-02-08 | 2018-09-07 | 주식회사 라디코 | Method, device and program for detecting an object |
| CN110781839A (en) * | 2019-10-29 | 2020-02-11 | 北京环境特性研究所 | Sliding window-based small and medium target identification method in large-size image |
Non-Patent Citations (4)
| Title |
|---|
| ALI SHARIFI BOROUJERDI等: "Deep Interactive Region Segmentation and Captioning" * |
| 李庆忠;李宜兵;牛炯;: "基于改进YOLO和迁移学习的水下鱼类目标实时检测" * |
| 李梁等: "改进Mask R-CNN在航拍灾害检测的应用研究" * |
| 王苹: "云计算模型下图像边缘重叠区域检测方法研究" * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132164A (en) * | 2020-11-20 | 2020-12-25 | 北京易真学思教育科技有限公司 | Target detection method, system, computer device and storage medium |
| CN112132164B (en) * | 2020-11-20 | 2021-03-09 | 北京易真学思教育科技有限公司 | Target detection method, system, computer device and storage medium |
| CN113762220A (en) * | 2021-11-03 | 2021-12-07 | 通号通信信息集团有限公司 | Object recognition method, electronic device, and computer-readable storage medium |
| WO2023116641A1 (en) * | 2021-12-21 | 2023-06-29 | 北京罗克维尔斯科技有限公司 | Parking space detection model training method and apparatus, and parking space detection method and apparatus |
| CN114332456A (en) * | 2022-03-16 | 2022-04-12 | 山东力聚机器人科技股份有限公司 | Target detection and identification method and device for large-resolution image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111191730B (en) | 2023-05-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111191730A (en) | Method and system for detecting oversized image target facing embedded deep learning | |
| US8744168B2 (en) | Target analysis apparatus, method and computer-readable medium | |
| CN100513997C (en) | Multiple angle movement target detection, positioning and aligning method | |
| CN108961235A (en) | A kind of disordered insulator recognition methods based on YOLOv3 network and particle filter algorithm | |
| CN111738206A (en) | Excavator detection method for unmanned aerial vehicle inspection based on CenterNet | |
| CN113902792B (en) | Building height detection method, system and electronic equipment based on improved RETINANET network | |
| CN114332233B (en) | Laser SLAM loop detection method and system | |
| CN114170230B (en) | Glass defect detection method and device based on deformable convolution and feature fusion | |
| CN113989604B (en) | Tire DOT information identification method based on end-to-end deep learning | |
| CN110598698A (en) | Natural scene text detection method and system based on adaptive regional suggestion network | |
| CN113205023B (en) | High-resolution image building extraction fine processing method based on prior vector guidance | |
| KR102285269B1 (en) | Image analysis apparatus and method for utilizing the big data base GEO AI | |
| CN112101113B (en) | Lightweight unmanned aerial vehicle image small target detection method | |
| CN110807430B (en) | Method for preprocessing live panoramic traffic sign picture | |
| CN107247967B (en) | Vehicle window annual inspection mark detection method based on R-CNN | |
| US20230021591A1 (en) | Model generation method, model generation apparatus, non-transitory storage medium, mobile object posture estimation method, and mobile object posture estimation apparatus | |
| CN112215073A (en) | Traffic marking line rapid identification and tracking method under high-speed motion scene | |
| WO2023070955A1 (en) | Method and apparatus for detecting tiny target in port operation area on basis of computer vision | |
| CN114862913A (en) | Machine vision target positioning method based on artificial intelligence network | |
| CN114608522A (en) | Vision-based obstacle identification and distance measurement method | |
| CN118015595A (en) | Vehicle target detection and identification method based on attention mechanism and dynamic convolution | |
| CN115457559B (en) | Method, device and equipment for intelligently correcting texts and license pictures | |
| JPH09245168A (en) | Picture recognizing device | |
| CN111860332B (en) | Dual-channel electrokinetic diagram part detection method based on multi-threshold cascade detector | |
| Li et al. | Depth image restoration method based on improved FMM algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |