CN-121982826-A - Mountain area large-scale animal intelligent monitoring system based on multi-light image fusion

CN121982826ACN 121982826 ACN121982826 ACN 121982826ACN-121982826-A

Abstract

The invention discloses an intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion, which relates to the technical field of interactive early warning and comprises a monitoring center, wherein the monitoring center is connected with a perception acquisition module, a fusion processing module, an early warning analysis module and an intelligent regulation and control module, acquires environmental temporal information of a patrol coverage area of operation and maintenance personnel, constructs a patrol holographic situation map, performs multi-light extraction on the environmental temporal information to obtain a visible light image and a thermal infrared image, performs double identification on the visible light image and the thermal infrared image, identifies a selected point species, performs animal positioning on the operation and maintenance personnel based on the selected point species to obtain animal threat distances, performs danger level assessment on the animal threat distances, generates a risk avoidance self-protection scheme of the operation and maintenance personnel according to an assessment result, realizes active early warning, strives for risk avoidance time, and has ultrahigh detection rate and extremely low false alarm rate.

Inventors

YUAN SHOUBIN
SUN YONG
SUN YANPING
ZHANG GUANGXIN
XU SHENGCHEN
WANG MINZHEN
GAO BO
Zhong Caiming
WU KUI

Assignees

国网吉林省电力有限公司延边供电公司
宁波大学科学技术学院

Dates

Publication Date: 20260505
Application Date: 20260203

Claims (8)

1. The intelligent monitoring system for the large animals in the mountainous area based on the multi-light image fusion comprises a monitoring center, and is characterized in that the monitoring center is connected with a perception acquisition module, a fusion processing module, an early warning analysis module and an intelligent regulation and control module; The perception acquisition module is used for acquiring environmental temporal information of the patrol coverage of operation and maintenance personnel; The fusion processing module is used for constructing a patrol holographic situation map according to a patrol coverage range, marking personnel perception nodes, performing multi-light extraction on environmental temporal information, and obtaining a visible light image and a thermal infrared image; The early warning analysis module is used for carrying out position mapping on the thermal infrared image to obtain thermal infrared image positioning, capturing a visible light image according to positioning to obtain a matched positioning light image, and carrying out species identification based on the thermal infrared image positioning to obtain a selected point species; the intelligent regulation and control module is used for carrying out animal positioning on operation and maintenance personnel based on selected point species, obtaining animal threat distances, carrying out danger level assessment according to the animal threat distances, obtaining perceived point threat levels, and generating a risk avoidance self-protection scheme of the operation and maintenance personnel through inspection holographic situation diagrams based on the perceived point threat levels.
2. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of collecting environmental temporal information comprises the following steps: Setting a multi-light acquisition end for operation and maintenance personnel, and carrying out patrol acquisition on the operation and maintenance personnel according to the obtained multi-light acquisition end to obtain a patrol coverage area; comprehensive collection is carried out on the patrol coverage through a multi-light collection end, so that environmental tense information is obtained.
3. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of constructing a patrol holographic situation map according to a patrol coverage comprises the following steps: acquiring a patrol coverage, and carrying out AR conversion on the patrol coverage to obtain a patrol holographic situation map; homomorphic conversion is carried out on operation and maintenance personnel based on the inspection holographic situation map, and personnel perception nodes are obtained; And marking the obtained personnel sensing nodes at the corresponding positions of the patrol holographic situation map.
4. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of obtaining visible light images and thermal infrared images comprises: Uploading the environmental temporal information to a patrol holographic situation map, and performing multi-light capturing on the environmental temporal information through the patrol holographic situation map to obtain light image data; Performing space-time registration on the obtained light image data to obtain calibration light map data; And performing light type separation on the obtained calibration light map data to obtain a visible light image and a thermal infrared image.
5. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of obtaining the matching positioning light map comprises the following steps: Acquiring a thermal infrared image, uploading the thermal infrared image to a patrol holographic situation map, performing position matching on the uploaded thermal infrared image to obtain thermal infrared image positioning, performing position marking on the thermal infrared image in the patrol holographic situation map, and marking the obtained thermal infrared image at a corresponding thermal infrared image positioning position; And carrying out matching screening on the visible light images based on the thermal red image positioning to obtain a matching positioning light map, and marking the matching positioning light map on the corresponding thermal red image positioning.
6. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of species identification based on thermal red image positioning comprises the following steps: setting animal heat red characteristic conditions; performing primary identification on the thermal red image positioning according to the obtained thermal red characteristic conditions of the animals to obtain a thermal red image of the large animals, and marking the thermal red image positioning corresponding to the thermal red image of the large animals as a primary selected positioning point; And (3) carrying out animal species identification on the obtained initial selected positioning points according to the large animal thermal red map by matching the positioning light map to obtain the selected point species.
7. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of obtaining the animal threat distance comprises: carrying out route simulation on personnel sensing nodes according to the selected point species through the patrol holographic situation map to obtain a species pre-escape route; Comparing the distances of the obtained species pre-escape routes to obtain a nearest pre-chase route; And carrying out distance statistics on the personnel sensing nodes according to the obtained nearest pre-chase route to obtain the animal threat distance.
8. The intelligent monitoring system for large animals in mountainous areas based on multi-light image fusion according to claim 1, wherein the process of generating the risk avoidance self-protection scheme of operation and maintenance personnel by inspecting a holographic situation map based on the threat level of a sensing point comprises the following steps: hazard classification is carried out on the threat distances of animals according to the selected point species, and threat distance grades are obtained; Performing basic threat grading on the selected point species to obtain an inherent threat grade; Classifying dynamic behaviors of the selected point species based on the environmental temporal information to obtain a behavior perception class; Performing risk assignment on the obtained threat distance level, the inherent threat level and the behavior perception level to obtain threat level points; Setting threat weights, and carrying out perceived point threat assessment on threat level scores according to the threat weights to obtain perceived point threat levels; and carrying out safety risk avoidance on the personnel sensing nodes according to the obtained sensing point threat level through the patrol holographic situation map to obtain a risk avoidance self-protection scheme.

Description

Mountain area large-scale animal intelligent monitoring system based on multi-light image fusion Technical Field The invention relates to the technical field of interactive early warning, in particular to a mountain area large-scale animal intelligent monitoring system based on multi-light image fusion. Background The monitoring of large animals in mountain areas is a key link of ecological protection, biodiversity research and conflict prevention and control of human beings and animals. The traditional monitoring means such as manual inspection, infrared triggering cameras and the like have obvious limitations that the manual inspection coverage area is small, the efficiency is low and the infrared monitoring means are limited by terrain and climate, the infrared cameras can automatically capture images, but are triggered by the body temperature of animals, the infrared monitoring means are easily interfered by the ambient temperature, and the single-spectrum imaging has limited recognition rate under the complex illumination condition. Especially in dense vegetation, night or bad weather, the details of the visible light image are lost, and the thermal infrared image can not provide texture features, so that species misjudgment or omission is caused. With the development of optical sensor technology, the multispectral imaging system is gradually applied to field monitoring, but the existing system still faces the challenges that the difficulty of collaborative calibration of the multispectral sensors is high, the hardware cost is limited to large-scale deployment, animal shapes are changeable, shielding is frequent, algorithm robustness is to be improved, and a mass monitoring data storage and analysis system is not complete. Therefore, the intelligent monitoring system for the large animals in the mountainous area is constructed efficiently and adaptively, a high-precision sensor network, a self-adaptive fusion algorithm and a cloud edge cooperative framework are required to be further integrated, all-weather and accurate monitoring of the large animals in the mountainous area is realized, reliable technical support is provided for ecological protection, and therefore the intelligent monitoring system for the large animals in the mountainous area based on multi-light image fusion is provided. Disclosure of Invention The invention aims to provide a mountain area large animal intelligent monitoring system based on multi-light image fusion, which aims to solve the problems of poor environmental adaptability and inaccurate early warning information in the background technology. The intelligent monitoring system for the large animals in the mountainous area based on the multi-light image fusion comprises a monitoring center, wherein the monitoring center is connected with a sensing acquisition module, a fusion processing module, an early warning analysis module and an intelligent regulation and control module; The perception acquisition module is used for acquiring environmental temporal information of the patrol coverage of operation and maintenance personnel; The fusion processing module is used for constructing a patrol holographic situation map according to a patrol coverage range, marking personnel perception nodes, performing multi-light extraction on environmental temporal information, and obtaining a visible light image and a thermal infrared image; The early warning analysis module is used for carrying out position mapping on the thermal infrared image to obtain thermal infrared image positioning, capturing a visible light image according to positioning to obtain a matched positioning light image, and carrying out species identification based on the thermal infrared image positioning to obtain a selected point species; the intelligent regulation and control module is used for carrying out animal positioning on operation and maintenance personnel based on selected point species, obtaining animal threat distances, carrying out danger level assessment according to the animal threat distances, obtaining perceived point threat levels, and generating a risk avoidance self-protection scheme of the operation and maintenance personnel through inspection holographic situation diagrams based on the perceived point threat levels. Preferably, the process of collecting the environmental temporal information includes: Setting a multi-light acquisition end for operation and maintenance personnel, and carrying out patrol acquisition on the operation and maintenance personnel according to the obtained multi-light acquisition end to obtain a patrol coverage area; comprehensive collection is carried out on the patrol coverage through a multi-light collection end, so that environmental tense information is obtained. Preferably, the process of constructing the patrol holographic situation map according to the patrol coverage comprises the following steps: acquiring a patrol coverage, and carrying out AR conversion on the