Elk identification and form contour parameter extraction method based on unmanned aerial vehicle remote sensing

Abstract

A method for elk recognition and form contour parameter extraction based on unmanned aerial vehicle remote sensing comprises the following steps: taking an unmanned aerial vehicle remote sensing image as a processing material, firstly performing preprocessing and labeling, then establishing a target detection model, and then training by using an initial image and a labeling file; substituting the new remote sensing image into the target detection model to obtain target detection position information of the elk individuals on the new remote sensing image, importing the target detection position information and the target detection position information into a database to obtain an image data set, screening to obtain an elk individual image semantic segmentation training set, generating an individual mask, and establishing a semantic segmentation model; the method comprises the following steps of: acquiring an individual mask, training by using a training set and the individual mask, calling a new image data set into a model to obtain an individual prediction mask, performing gray scale and binarization processing, and converting a binarized graph into an elk contour vector information file to complete morphological contour parameter extraction. According to the design, the monitoring precision is high in a complex environment, elk form contour parameter extraction is facilitated, the use cost is low, and the functionality is high.

Description

technical field

[0001] The invention relates to a monitoring method for elk, belonging to the field of animal monitoring, in particular to a method for elk identification and morphological contour parameter extraction based on remote sensing of unmanned aerial vehicles. Background technique

[0002] Elk is a large-scale ungulate wetland species unique to China, and it is a national-level key protected wild animal. Due to human activities, climate change and other factors, wild elk populations became extinct in mainland China in the early 19th century. In 1985, China began to implement the elk reintroduction project. At present, the number of elk populations in my country has exceeded 8,000, and three large wild elk populations have gradually formed in Dafeng in Jiangsu, Shishou in Hubei and Dongting Lake in Hunan. small population. In 1991, Hubei Shishou Milu National Nature Reserve was established on the old Yangtze River Road in Swanzhou in the middle reaches of the Yangt...

Examples

Embodiment 1

[0078] see figure 1 , a method for elk identification and morphological profile parameter extraction based on UAV remote sensing, comprising the following steps:

[0079] S1: Photograph the elk population on the ground through the camera or sensor mounted on the UAV platform to obtain the corresponding UAV remote sensing image. The UAV remote sensing image is a picture, and the number of shooting times is multiple. The number of remote sensing images is multiple;

[0080] S2: Preprocess the aforementioned UAV remote sensing image to obtain the initial image I1 (such as figure 2 shown), the preprocessing includes at least one or any combination of image enhancement, denoising, and filtering; all initial images I1 form a separate folder;

[0081] S3: Use the target recognition data labeling tool to label the elk individuals in the initial image I1 (such as image 3 After each initial image I1 is marked, a marked file T1 is correspondingly generated, and the T1 is in txt or x...

Embodiment 2

[0091] The basic content is the same as that of Example 1, except that:

[0092] like Figure 10 As shown, in S9, the OpenCV library performs grayscale processing and binarization processing on the individual prediction masks in turn; at the same time, the OpenCV library calls the binarized graphics, and superimposes the binarized graphics on its corresponding On the individual pictures of elk in the new image dataset D2, and display it.

Embodiment 3

[0094] The basic content is the same as that of Example 1, except that:

[0095] like Figure 11 , Figure 12 As shown, first use ArcGIS tool to convert the above binarized graphics into elk outline vector information file, and then obtain the area and perimeter of elk outline based on the elk outline vector information file, and generate skeleton lines, and then calculate the length of the extracted skeleton lines as The length of the elk, and then determine the sex of the elk according to the ratio of the length of the elk to the circumference of the outline;

[0096] In addition, first extract feature points from the outline of the elk, and then calculate according to the feature points to obtain the head width and body width of the elk, and then determine the age of the elk according to the body length and body width;

[0097] Finally, after the above judgment operation is completed, the obtained area, perimeter, body length, head width, body width, gender, and age are i...

technical field

[0001] The invention relates to a monitoring method for elk, belonging to the field of animal monitoring, in particular to a method for elk identification and morphological contour parameter extraction based on remote sensing of unmanned aerial vehicles. Background technique

[0002] Elk is a large-scale ungulate wetland species unique to China, and it is a national-level key protected wild animal. Due to human activities, climate change and other factors, wild elk populations became extinct in mainland China in the early 19th century. In 1985, China began to implement the elk reintroduction project. At present, the number of elk populations in my country has exceeded 8,000, and three large wild elk populations have gradually formed in Dafeng in Jiangsu, Shishou in Hubei and Dongting Lake in Hunan. small population. In 1991, Hubei Shishou Milu National Nature Reserve was established on the old Yangtze River Road in Swanzhou in the middle reaches of the Yangt...

Examples

Embodiment 1

[0078] see figure 1 , a method for elk identification and morphological profile parameter extraction based on UAV remote sensing, comprising the following steps:

[0079] S1: Photograph the elk population on the ground through the camera or sensor mounted on the UAV platform to obtain the corresponding UAV remote sensing image. The UAV remote sensing image is a picture, and the number of shooting times is multiple. The number of remote sensing images is multiple;

[0080] S2: Preprocess the aforementioned UAV remote sensing image to obtain the initial image I1 (such as figure 2 shown), the preprocessing includes at least one or any combination of image enhancement, denoising, and filtering; all initial images I1 form a separate folder;

[0081] S3: Use the target recognition data labeling tool to label the elk individuals in the initial image I1 (such as image 3 After each initial image I1 is marked, a marked file T1 is correspondingly generated, and the T1 is in txt or x...

Embodiment 2

[0091] The basic content is the same as that of Example 1, except that:

[0092] like Figure 10 As shown, in S9, the OpenCV library performs grayscale processing and binarization processing on the individual prediction masks in turn; at the same time, the OpenCV library calls the binarized graphics, and superimposes the binarized graphics on its corresponding On the individual pictures of elk in the new image dataset D2, and display it.

Embodiment 3

[0094] The basic content is the same as that of Example 1, except that:

[0095] like Figure 11 , Figure 12 As shown, first use ArcGIS tool to convert the above binarized graphics into elk outline vector information file, and then obtain the area and perimeter of elk outline based on the elk outline vector information file, and generate skeleton lines, and then calculate the length of the extracted skeleton lines as The length of the elk, and then determine the sex of the elk according to the ratio of the length of the elk to the circumference of the outline;

[0096] In addition, first extract feature points from the outline of the elk, and then calculate according to the feature points to obtain the head width and body width of the elk, and then determine the age of the elk according to the body length and body width;

[0097] Finally, after the above judgment operation is completed, the obtained area, perimeter, body length, head width, body width, gender, and age are i...

Images