Shulou(Shulou.com)06/03 Report--

This article introduces the relevant knowledge of "how to realize object detection and tracking". In the operation of actual cases, many people will encounter such a dilemma, so let the editor lead you to learn how to deal with these situations. I hope you can read it carefully and be able to achieve something!

Effect exampl

Basic principles and implementation steps

Implementation steps

Image preprocessing: in the process of image preprocessing, through the basic operations such as coding, threshold and filtering, improved mode, discrete mode operation, etc., on the basis of keeping all the content information of the image as much as possible, the image features are sorted out and highlighted in order to achieve the best state of image feature extraction.

Image feature extraction: in the process of feature extraction, the pre-processed image can extract high-value features, so that the high-dimensional feature space has better separation and is more convenient for algorithm recognition. Image feature extraction is generally subdivided into several modules: basic visual feature extraction such as grayscale value, luminance value, shape and texture, color feature extraction based on pixel color, texture feature extraction of object texture and structural material, shape feature extraction based on contour and region, and spatial feature extraction of spatial position relationship information between multiple objects segmented from the image.

Feature selection: after feature extraction, feature selection can be carried out according to the specific object recognition. For example, in the case of a wide variety of features and many kinds of objects, we can find the occasions where each feature fits through feature selection.

Modeling: after feature selection, the feature set is established in the modeling process, the similarities and differences are distinguished, and the similarities and differences are extracted. The main modeling object is the spatial structure relationship between features.

Matching: the matching step is to use the model to identify and match the new image, to identify what kind of object the image belongs to, and the condition can separate the object from the rest of the image.

Positioning: after object recognition, the target object is located, that is, the identified object coordinates are combined with the actual physical space coordinates, and then the corresponding tracking and model superposition processing can be carried out.

Three capabilities provided by AR Engine

The three capabilities provided by Huawei AR Engine are motion tracking, environment tracking, human body tracking and face tracking.

Motion tracking is mainly through the terminal equipment camera to identify feature points, and track the movement changes of these feature points, to constantly track the position and attitude of the terminal equipment.

Environment tracking can identify the plane, such as the ground, walls, etc., and can also estimate the light intensity around the plane.

Human body and face tracking make terminal devices have the ability to understand people. By locating the position of the human hand and recognizing specific gestures, the special effects of virtual objects or contents can be placed on the human hand; combined with depth devices, the motion tracking of 21 skeletal points of the hand can be accurately restored to do more refined interactive control and superposition of special effects. When the recognition range is extended to the whole body, the 23 key positions of the human body can be used to detect the posture of the human body in real time, which provides ability support for the application development of somatosensory and sports health.

Development preparation

Add permissions in AndroidManifest.xml

Open the AndroidManifest.xml file in main, and in the

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.