Target recognition is an intelligent video-analysis technology that identifies or matches a target by extracting and comparing distinguishing visual features from captured images.
Option B is correct because target recognition can be applied to persons appearing in surveillance video. After a target has been detected, the recognition system locates relevant feature points, extracts distinguishing information, and compares the resulting features with reference images or feature records.
Option A is incorrect because the recognition process requires image information. This is ordinarily obtained from cameras or existing image and video sources. Without image collection or access to previously collected images, the described visual recognition workflow cannot operate.
Option C is incorrect because target matching is not inherently or predominantly performed by cameras. Intelligent cameras can support edge-side target analysis, but matching can also be performed by centralized intelligent-analysis services or platform components. The question's broad assertion about where most matching occurs is not established by the Huawei architecture.
Option D is incorrect because options A and C are incorrect.
Huawei's course identifies image collection, target detection, feature-point location, feature extraction, feature comparison, and result output as the essential stages of recognition.
The laboratory guide also distinguishes target detection and capture from target matching, emphasizing their related but separate functions.
Official Huawei references:
Training Material — Target Recognition, slide 97. Lab Guide — Experiment Environment Overview, pages 6–7; Experiment 7, Target and Person Detection.