CN-121995374-A - Target detection method, system, equipment and storage medium based on ground penetrating radar

CN121995374ACN 121995374 ACN121995374 ACN 121995374ACN-121995374-A

Abstract

The application belongs to the field of target detection, and in particular relates to a target detection method, a system, equipment and a storage medium based on a ground penetrating radar, wherein the method comprises the steps of using the ground penetrating radar to run along a straight line, detecting a target to be detected in the running process, and acquiring echo data of the target to be detected at different detection points; normalizing echo data to obtain a normalized image, converting the normalized image into a gray level image to obtain a gray level image, dividing the gray level image into a plurality of gray level subgraphs, calculating a gray level histogram of each gray level subgraph, carrying out contrast enhancement on the gray level image according to the gray level histogram to obtain an enhanced image, carrying out size and pixel adjustment on the enhanced image to obtain an adjusted image, inputting the adjusted image into a pre-trained YOLO model to obtain the type, the position and the confidence of a target to be detected, and taking the type, the position and the confidence of the target as detection results. The application has the effects of improving the detection accuracy and reducing the omission factor.

Inventors

WANG MENG
CHANG YULIANG
HOU WEIBING
CHU XIANG

Assignees

湖南艾科诺维科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260122

Claims (10)

1. The target detection method based on the ground penetrating radar is characterized by comprising the following steps of: the method comprises the steps of using a ground penetrating radar to run along a straight line, detecting a target to be detected in the running process, and obtaining echo data of the target to be detected at different detection points; Normalizing the echo data to obtain a normalized image, and converting the normalized image into a gray level image to obtain a gray level image; dividing the gray level image into a plurality of gray level subgraphs; Obtaining the occurrence frequency of each gray level according to the gray level subgraph, and calculating according to the occurrence frequency of the gray level to obtain an accumulated distribution value; If the accumulated distribution value is larger than a preset threshold value, clipping the gray sub-graph, uniformly distributing the gray level with the accumulated distribution value larger than the preset threshold value to other gray levels after gray level adjustment, and calculating the accumulated distribution value of the adjusted gray sub-graph until no accumulated distribution value is larger than the preset threshold value; normalizing the cumulative distribution value of the adjusted gray subgraph to obtain a normalized value; mapping is carried out according to the normalization value so as to realize the contrast enhancement of the gray level image and obtain an enhanced image; Performing size and pixel adjustment on the enhanced image to obtain an adjusted image; and inputting the adjustment image into a pre-trained YOLO model to obtain the target type, the position and the confidence of the target to be detected, and taking the target type, the position and the confidence as detection results.
2. The method for detecting a target based on a ground penetrating radar according to claim 1, wherein normalizing the echo data to obtain normalized data, converting the normalized data into a gray scale map to obtain a gray scale image, comprises: Representing the echo data as radar echo signals, as: Wherein, the Is the first The coordinates of the individual detection points in the detection plane, For the total number of probe points, In order to be able to take time, And Respectively the ground penetrating radars are at the first The amplitude and phase of the echo are received by the detection points; drawing a two-dimensional image according to the radar echo signal; Normalizing the two-dimensional image to obtain a normalized image; and converting the normalized image into a gray scale image to obtain a gray scale image.
3. The method for detecting a target based on a ground penetrating radar according to claim 1, wherein the obtaining the number of occurrences of each gray level based on the gray level subgraph, and calculating the cumulative distribution value based on the number of occurrences of the gray level, comprises: counting the occurrence times of each gray level according to the gray subgraph; and obtaining a cumulative distribution value according to the occurrence times of the gray levels and the total number of pixels of the gray sub-images.
4. The ground penetrating radar based target detection method of claim 1, wherein the calculating the cumulative distribution value of the adjusted gray subgraph comprises: calculating to obtain the number of residual pixels after clipping according to the occurrence times and the clipping threshold; Calculating to obtain the number of occurrences after clipping according to the number of the residual pixels and a clipping threshold; and calculating according to the number of occurrences after clipping to obtain the cumulative distribution value of the adjusted gray subgraph.
5. The method for detecting a target based on ground penetrating radar according to claim 1, wherein the normalization of the cumulative distribution value of the adjusted gray subgraph is performed to obtain a normalized value expressed as: Wherein, the , For the cumulative distribution value of the adjusted gray subgraph, Is a normalized value.
6. The method for detecting a target based on ground penetrating radar according to claim 1, wherein the inputting the adjustment image into a pre-trained YOLO model to obtain a target type, a position and a confidence level of the target to be detected comprises: Extracting the characteristics to be predicted according to the adjustment image; and inputting the characteristics to be predicted into a pre-trained YOLO model to obtain the target type, the position and the confidence of the target to be detected.
7. A ground penetrating radar-based target detection system, comprising: The data acquisition module is used for using the ground penetrating radar to run along a straight line, detecting the target to be detected in the running process, and acquiring echo data of the target to be detected at different detection points; the image processing module is used for normalizing the echo data to obtain a normalized image, and converting the normalized image into a gray level image to obtain a gray level image; The splitting module is used for dividing the gray image into a plurality of gray subgraphs; The calculation module is used for obtaining the occurrence frequency of each gray level according to the gray level subgraph and obtaining an accumulated distribution value according to the occurrence frequency of the gray level; The adjustment module is used for cutting the gray subgraph if the accumulated distribution value is larger than a preset threshold value, uniformly distributing the gray level with the accumulated distribution value larger than the preset threshold value to other gray levels after gray level adjustment, and calculating the accumulated distribution value of the adjusted gray subgraph until no accumulated distribution value is larger than the preset threshold value; the normalization module is used for normalizing the cumulative distribution value of the adjusted gray subgraph to obtain a normalization value; the enhancement module is used for mapping according to the normalization value so as to realize contrast enhancement of the gray level image and obtain an enhanced image; the image adjusting module is used for adjusting the size and the pixels of the enhanced image to obtain an adjusted image; The detection module is used for inputting the adjustment image into a pre-trained YOLO model to obtain the target type, the position and the confidence coefficient of the target to be detected, and the target type, the position and the confidence coefficient are used as detection results.
8. The ground penetrating radar-based target detection system of claim 1, wherein the computing module comprises: The statistics unit is used for counting the occurrence times of each gray level according to the gray subgraph; and the calculating unit is used for obtaining a cumulative distribution value according to the occurrence times of the gray levels and the total number of pixels of the gray sub-image.
9. A terminal device comprising a memory and a processor, characterized in that the memory stores a computer program capable of running on the processor, which processor, when loaded and executed, employs the method according to any of claims 1-7.
10. A computer readable storage medium having a computer program stored therein, which, when loaded and executed by a processor, employs the method of any of claims 1 to 7.

Description

Target detection method, system, equipment and storage medium based on ground penetrating radar Technical Field The invention belongs to the field of target detection, and particularly relates to a target detection method, a target detection system, target detection equipment and a target detection storage medium based on a ground penetrating radar. Background The ground penetrating radar is widely applied to the fields of urban road disease detection (such as hole identification), underground pipeline detection, archaeological exploration and the like. The method is characterized in that positioning is realized by analyzing hyperbolic characteristics of target echoes (such as metal pipelines, cracks and hollows) in Bscan echo images. However, the problems of low signal-to-noise ratio, weak target characteristics, poor image quality and the like make Bscan images insufficient in contrast, and it is difficult to highlight hyperbolic characteristics. The existing ground penetrating radar adopts enhancement methods such as linear lifting, wavelet denoising, global histogram homogenization and the like, and the methods are difficult to effectively inhibit noise while keeping weak signals, so that the follow-up target detection omission ratio is higher. In the aspect of detection, although the YOLO series model based on deep learning is excellent in general target detection in recent years, the YOLO series model still faces challenges in the field of ground penetrating radars, namely, firstly, the size of a target echo of the ground penetrating radars is small, a YOLO default Anchor Box (Anchor Box) is difficult to match, secondly, the original Bscan images have large noise interference and low contrast, and the direct input of the model is easy to cause false detection. Disclosure of Invention The invention aims to solve the technical problem of providing a target detection method, a system, equipment and a storage medium based on a ground penetrating radar, which are used for improving the target detection accuracy and reducing the omission ratio by carrying out contrast enhancement on images. A target detection method based on ground penetrating radar comprises the following steps: the method comprises the steps of using a ground penetrating radar to run along a straight line, detecting a target to be detected in the running process, and obtaining echo data of the target to be detected at different detection points; Normalizing the echo data to obtain a normalized image, and converting the normalized image into a gray level image to obtain a gray level image; dividing the gray level image into a plurality of gray level subgraphs; Obtaining the occurrence frequency of each gray level according to the gray level subgraph, and calculating according to the occurrence frequency of the gray level to obtain an accumulated distribution value; If the accumulated distribution value is larger than a preset threshold value, clipping the gray sub-graph, uniformly distributing the gray level with the accumulated distribution value larger than the preset threshold value to other gray levels after gray level adjustment, and calculating the accumulated distribution value of the adjusted gray sub-graph until no accumulated distribution value is larger than the preset threshold value; normalizing the cumulative distribution value of the adjusted gray subgraph to obtain a normalized value; mapping is carried out according to the normalization value so as to realize the contrast enhancement of the gray level image and obtain an enhanced image; Performing size and pixel adjustment on the enhanced image to obtain an adjusted image; and inputting the adjustment image into a pre-trained YOLO model to obtain the target type, the position and the confidence of the target to be detected, and taking the target type, the position and the confidence as detection results. Optionally, normalizing the echo data to obtain normalized data, and converting the normalized data into a gray scale map to obtain a gray scale image, including: Representing the echo data as radar echo signals, as: Wherein, the Is the firstThe coordinates of the individual detection points in the detection plane,For the total number of probe points,In order to be able to take time,AndRespectively the ground penetrating radars are at the firstThe amplitude and phase of the echo are received by the detection points; drawing a two-dimensional image according to the radar echo signal; Normalizing the two-dimensional image to obtain a normalized image; and converting the normalized image into a gray scale image to obtain a gray scale image. Optionally, the obtaining the occurrence number of each gray level according to the gray level subgraph, and calculating to obtain the cumulative distribution value according to the occurrence number of the gray level, includes: counting the occurrence times of each gray level according to the gray subgraph; and obtaining a cumulative dis