智能检测打火机算法模型的X光图像训练数据

该数据知识产权可应用在各类安检场景中，可精准、快速检测被检物品中是否包含打火机物品。1、数据来源：应用X射线光源多维度、多角度、多场景下透射多种类型的打火机采集并建立其原始的X光数据图例库。 2、数据深度处理：对采集到的原始X光图像预标注坐标位置和品项类别，并对打火机图像进行抠图处理。将抠出的打火机图像与多场景的图像分别进行几何变换、像素变换等增广处理。 3、检测模型生成算法规则：将处理后的打火机X光图像和场景图像通过密度统计（像素值代表实物密度值）依据区域匹配原则进行融合，融合区域掩模作为数据标签与融合后的图像作为深度学习样本数据。还可通过调整抠图区域在场景图像区域的位置，获得不同的平均密度差值，训练生成可精准定位、精准识别打火机的智能检测模型。区域匹配原则按照Mask*(α*ρ抠图图像+β*ρ场景图像)，融合后的图像处理公式按照Mask*(α*ρ抠图图像+β*ρ抠图图像)+（1-Mask）*ρ场景图像。（所述公式中：Mask为图像掩膜，图像目标区域值为1，目标区域外值为0，ρ为密度值，α、β指系数）检测模型可对多场景下、形态各异的打火机精准识别，同时将目标物的位置及所在X光图像信息记录标出。进一步的还可根据目标物位置信息推算目标物尺寸信息。

This dataset’s intellectual property can be applied to various security inspection scenarios, enabling accurate and rapid detection of lighters in inspected articles.
1. Data Source: An original X-ray image database of various types of lighters is established by collecting transmission X-ray data of lighters under multi-dimensional, multi-angle and multi-scenario conditions using X-ray sources.
2. In-depth Data Processing: Pre-label the coordinate positions and item categories of the collected original X-ray images, and perform matting processing on the lighter images. Then, perform augmentation processing including geometric transformation and pixel transformation on the extracted lighter images and multi-scenario images respectively.
3. Algorithm Rules for Generating Detection Models: Fuse the processed X-ray images of lighters and scenario images based on the region matching principle through density statistics (pixel values represent the density values of physical objects). The fused region mask is used as the data label, and the fused image is used as the deep learning sample data. Additionally, by adjusting the position of the matting area within the scenario image area, different average density differences can be obtained to train and generate an intelligent detection model that can accurately locate and identify lighters. The region matching principle follows the formula: Mask*(α*ρ_matting + β*ρ_scenario), and the image processing formula for the fused image follows: Mask*(α*ρ_matting + β*ρ_matting) + (1-Mask)*ρ_scenario. (In the formulas mentioned above: Mask is the image mask, with a value of 1 in the target region of the image and 0 outside the target region; ρ represents the density value, and α and β refer to coefficients.)
The detection model can accurately identify lighters of various shapes in multiple scenarios, while recording and marking the position of the target object and the relevant information of the X-ray image it is located in. Furthermore, the size information of the target object can also be inferred based on its position information.