Handmade model surface processing method based on texture mapping

Abstract

The invention relates to the technical field of handmade model surface processing, in particular to a handmade model surface processing method based on texture mapping, which comprises the steps of collecting three-dimensional surface data to generate an initial texture mapping graph, performing texture reconstruction on a missing or fuzzy region through an intelligent algorithm and a deep learning model, and obtaining a texture reconstruction result. And a multi-layer spraying process is combined with real-time monitoring to optimize spraying parameters, and finally, the quality is ensured through optical scanning and local repairing. The reality sense and production efficiency of the surface texture of the handmade model can be remarkably improved, meanwhile, the process complexity and cost are reduced, and the efficient and high-quality requirements of the market are met.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of digital image processing and three-dimensional model manufacturing, and specifically relates to a hand-made model surface processing method based on texture mapping. BACKGROUND

[0002] With the continuous development of hand-made model manufacturing technology, users' requirements for hand-made model surface processing effects are increasingly improved, especially in terms of texture performance, realism and production efficiency. At present, the surface processing technology of hand-made models has made certain progress, but still has many deficiencies. For example, the patent with the publication number CN114316842B (published on May 26, 2023) proposes a spraying glue for hand-made models and a preparation method and spraying process thereof, which sprays the hand-made model with a specific proportion of spraying glue to make the surface of the hand-made model present a gloss, texture and touch similar to real skin. However, this technical solution mainly relies on the physical properties of the spraying material to simulate the surface effect, lacks precise mapping capability for complex textures, and is difficult to meet the application scenarios of high-precision texture requirements. In addition, the spraying process has high requirements for the operating environment and equipment, which may increase the production cost and process complexity, limiting its application in efficient production.

[0003] On the other hand, the patent with the publication number CN117934867B (published on May 31, 2024) proposes a hand-made model production method and system based on intelligent adjustment function, which accurately colors the surface defects of the hand-made model by intelligently identifying and adjusting the coloring parameters, thereby improving the overall quality and production efficiency of the hand-made model. However, this technical solution focuses on the intelligent repair of coloring defects and does not involve the generation and mapping of complex textures, which cannot achieve high restoration of the surface texture of the hand-made model. At the same time, this system is highly dependent on image recognition and knowledge graph, which may limit the accuracy of texture mapping due to insufficient data accumulation or algorithm limitations, further affecting the realism and consistency of the surface processing effect.

[0004] The above-mentioned existing technologies show that the existing hand-made model surface processing technology still has significant deficiencies in precise mapping of complex textures, realistic restoration of surface texture and process optimization. Especially in application scenarios with high-precision texture requirements, the existing technology is difficult to balance the realism of texture performance, the simplification of process flow and the reduction of production cost. Therefore, an innovative hand-made model surface processing method is urgently needed to realize rapid mapping and realistic restoration of high-precision textures, optimize the process flow and reduce the production cost, so as to meet the market demand for efficient and high-quality surface processing of hand-made models. SUMMARY

[0005] The application provides a hand model surface processing method based on texture mapping, which aims to solve the defects of the prior art, such as insufficient precision mapping ability of complex texture on the surface of a hand model, low production efficiency and high cost. The hand model surface processing method comprises the following steps: collecting three-dimensional surface data of a hand model to be processed, and generating a high-precision initial texture mapping graph based on the three-dimensional surface data; when there is a texture missing or fuzzy area in the initial texture mapping graph, the geometric feature parameters of the area are extracted by an intelligent algorithm, and the matching degree with a preset standard texture library is calculated; when the matching degree is lower than a first preset threshold, the texture of the area is reconstructed based on a deep learning model, and the texture reconstruction is calculated by the following formula:

[0006]

[0007] wherein T new represents the texture value after reconstruction, T base represents the basic texture value, T ref,i represents the i th reference texture value, w i represents the weight coefficient of the i th reference texture, and a represents the retention ratio of the basic texture and satisfies 0≤a≤1. The formula dynamically adjusts the weight coefficient and the retention ratio to ensure the authenticity and consistency of the texture reconstruction.

[0008] If the area after texture reconstruction meets the first preset quality condition, the reconstructed texture is mapped to the surface of the hand model, and a multi-layer spraying process is started; during the multi-layer spraying process, the spraying thickness and uniformity are monitored in real time, and the spraying parameters, including the spraying pressure P and the spraying speed V, are adjusted based on the monitoring results, and the adjustment formula is as follows:

[0009] P adj =P init ·(1+k1·Δh),V adj =V init ·(1-k2·Δu)

[0010] wherein P adj and V adj represent the adjusted spraying pressure and speed, P init and V init represent the initial spraying pressure and speed, Δh represents the spraying thickness deviation, Δu represents the spraying uniformity deviation, and k1 and k2 are adjustment coefficients. Through the above formula, the intelligent optimization of the spraying process can be realized, and the surface treatment effect is improved.

[0011] After completing multi-layer spraying, the final surface image of the hand model is obtained based on optical scanning technology, and the surface texture distribution characteristics are extracted through image segmentation algorithm; whether the second preset quality condition is met is judged according to the texture distribution characteristics, if not, the specific area is processed again by using local repair process until the quality requirement is met. Through the collaborative optimization of texture mapping and spraying process, the reality and production efficiency of the surface texture of the hand model are significantly improved, and the process complexity and cost are reduced, which meets the market demand for efficient and high-quality surface treatment of the hand model. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 : The flowchart of the method of the present application shows the overall steps from three-dimensional surface data acquisition to the completion of the final surface treatment, including texture mapping, texture reconstruction, multi-layer spraying and quality detection, etc.

[0013] Figure 2 : The principle diagram of texture reconstruction and spraying parameter adjustment, which details the relationship between each parameter in the texture reconstruction formula and the calculation process of dynamic adjustment of spraying pressure and speed.

[0014] Figure 3 : The optical scanning and local repair process diagram shows the process of image segmentation analysis after obtaining the surface image of the hand model by optical scanning, and implementing local repair on the area that does not meet the quality condition. DETAILED DESCRIPTION

[0015] The present application provides a hand model surface treatment method based on texture mapping, which is combined with the Figure 1 , Figure 2 and Figure 3 , and the specific embodiments are described in detail. In this embodiment, the surface treatment of a high-precision hand model is taken as an example to introduce the entire process from three-dimensional surface data acquisition to the completion of the final surface treatment, including texture mapping, texture reconstruction, multi-layer spraying and quality detection, etc.

[0016] Firstly, as shown in Figure 1 , the first step of the method is to collect the three-dimensional surface data of the hand model to be processed. In actual operation, a high-precision three-dimensional scanner is used to scan the target hand model in all directions to obtain its complete three-dimensional point cloud data. In order to ensure the accuracy of the data, appropriate scanning resolution and angle range should be set during the scanning process, and the data integrity should be optimized by multiple scanning superposition. Then, the computer aided design (CAD) software is used to process the collected three-dimensional point cloud data to generate a high-precision initial texture mapping diagram. The initial texture mapping diagram contains the basic texture information of the surface of the hand model, which provides a basis for subsequent texture analysis and processing.

[0017] When the initial texture map is generated, it needs to be evaluated for quality. As shown in Figure 1 the evaluation process, if it is found that there are texture missing or blurred areas in the initial texture map, it enters the texture reconstruction stage. Specifically, the geometric feature parameters of these areas are extracted by intelligent algorithm, mainly including curvature, normal vector direction and boundary contour information. Then, the extracted geometric feature parameters are matched with the texture samples in the preset standard texture library. The matching degree calculation formula is as follows:

[0018]

[0019] Where M represents the matching degree, S i represents the similarity score of the i-th texture sample, w i represents the weight coefficient of the i-th texture sample. Through the above formula, the similarity of the current area and each texture sample in the standard texture library can be quantified. If the matching degree is lower than the first preset threshold (for example, 0.7), it is determined that the area needs to be reconstructed. The texture reconstruction process is as shown in Figure 2 , which uses a deep learning model to reconstruct the texture missing or blurred area. Specifically, the texture reconstruction formula is as follows:

[0020]

[0021] Where T new represents the reconstructed texture value, T base represents the base texture value, T ref,i represents the i-th reference texture value, w i represents the weight coefficient of the i-th reference texture, and a represents the retention ratio of the base texture and satisfies 0≤a≤1. In practical applications, the base texture value comes from the initial texture map of the hand model to be processed, and the reference texture value is selected from the standard texture library. The weight coefficient w i is calculated according to the geometric feature similarity of the reference texture and the target area. The higher the similarity, the greater the weight coefficient. By dynamically adjusting a and w i , it can ensure that the reconstructed texture not only maintains consistency with the original texture, but also has realistic and detailed performance.

[0022] After the texture reconstruction is completed, the reconstructed area needs to be evaluated for quality. The evaluation criteria include texture clarity, color consistency, and transition naturalness with the surrounding area. If the reconstructed area meets the first preset quality condition, it is mapped to the surface of the hand model, and the multi-layer spraying process is started. The multi-layer spraying process is as shown in Figure 1 , which is one of the core links of the method, and its purpose is to achieve efficient and high-quality surface treatment effect through intelligent spraying parameter adjustment.

[0023] In the multi-layer spraying process, real-time monitoring of the spraying thickness and uniformity is a critical step. Specifically, ultrasonic sensors and optical sensors are used to monitor the spraying thickness deviation Δh and spraying uniformity deviation Δu, respectively. Based on the monitoring results, the spraying pressure P and spraying speed V are dynamically adjusted, with the adjustment formulas as follows:

[0024] P adj =P init ·(1+k1·Δh),V adj =V init ·(1-k2·Δu)

[0025] where P adj and V adj represent the adjusted spraying pressure and speed, P init and V init represent the initial spraying pressure and speed, Δh represents the spraying thickness deviation, Δu represents the spraying uniformity deviation, and k1 and k2 are adjustment coefficients. In practical applications, the adjustment coefficients k1 and k2 typically range from 0.1 to 0.5, with specific values determined based on the characteristics of the spraying material and process requirements. Through the above formulas, real-time optimization of spraying parameters can be achieved, thereby improving spraying quality and efficiency.

[0026] After completing the multi-layer spraying, as shown in Figure 3 , an optical scanning technique is used to obtain the final surface image of the hand model. The optical scanning device captures the micro details of the hand model surface through laser scanning or structured light scanning, and generates a high-resolution digital image. Subsequently, an image segmentation algorithm is used to process the surface image and extract surface texture distribution features. The image segmentation algorithm uses an edge detection and region growing-based method, which can effectively distinguish different texture regions and extract their feature parameters.

[0027] Based on the extracted texture distribution features, it is determined whether the surface of the hand model meets the second preset quality condition. The second preset quality condition includes texture continuity, color saturation, and surface smoothness, among other indicators. If the quality condition is not met, a local repair process is used to perform secondary processing on specific areas. The local repair process, as shown in Figure 3 , mainly includes the following steps: first, a high-precision airbrush is used to perform local spraying on the surface of the hand model, with the spraying material selected to be the same as the original spraying material; second, a micro polishing tool is used to finely polish the sprayed area to ensure a smooth surface without marks; and finally, optical scanning and quality assessment are performed again until the quality requirements are met.

[0028] In this embodiment, a hand-made model of an animation character is taken as an example to describe the complete process from three-dimensional surface data acquisition to final surface treatment. In actual production, this method significantly improves the realism and production efficiency of the surface texture of the hand-made model, while reducing the complexity and cost of the process. For example, through intelligent texture reconstruction and spray parameter adjustment, the surface treatment time of a single hand-made model is shortened by about 30%, while the surface treatment quality is improved by about 20%. In addition, this method also has strong universality and can be widely applied to the surface treatment field of other complex curved surface models, such as sculptures, handicrafts, etc.

[0029] In summary, the specific embodiments of the present application combine the attached Figure 1 to the attached Figure 3 , each step of the hand-made model surface treatment method based on texture mapping and its technical principles are described in detail. From three-dimensional surface data acquisition to final surface treatment, the entire process covers texture mapping, texture reconstruction, multi-layer spraying and quality detection, etc. Through intelligent algorithms and dynamic adjustment mechanism, high-efficiency and high-quality surface treatment effect is achieved. This method not only solves the problem of insufficient precision mapping of complex surface texture of hand-made model in the prior art, but also significantly improves the production efficiency and reduces the cost, providing important technical support for the hand-made model manufacturing industry.

Claims

1. A surface treatment method for a figurine model based on texture mapping, characterized in that: The surface treatment method for a figure model comprises: collecting three-dimensional surface data of a figure model to be processed, and generating a high-precision initial texture map based on the three-dimensional surface data; when there are texture missing or blurred areas in the initial texture map, extracting geometric feature parameters of the area using an intelligent algorithm, and calculating the degree of matching between the area and a preset standard texture library; when the matching degree is lower than a first preset threshold, reconstructing the texture of the area based on a deep learning model, and calculating the texture reconstruction using the following formula: Among them, T new Represents the reconstructed texture value, T base Represents the base texture value, T ref,i represents the i-th reference texture value, w i represents the weight coefficient of the i-th reference texture, ɑ represents the retention ratio of the basic texture and satisfies 0≤ɑ≤1.

2. The surface treatment method of a figurine model according to claim 1, characterized in that: If the texture reconstructed area meets the first preset quality condition, the reconstructed texture is mapped to the surface of the figurine model and a multi-layer spraying process is started.

3. The surface treatment method of a figurine model according to claim 2, characterized in that: During the multi-layer spraying process, the spraying thickness and uniformity are monitored in real time, and the spraying pressure and spraying speed are adjusted based on the monitoring results. The adjustment formulas for the spraying pressure and spraying speed are as follows: P adj =P init ·(1+k1·Δh),V adj =V init ·(1-k2·Δu) Among them, P adj and V adj Represent the adjusted spraying pressure and speed, P init and V init They represent the initial spraying pressure and speed respectively, Δh represents the spraying thickness deviation, Δu represents the spraying uniformity deviation, and k1 and k2 are adjustment coefficients.

4. The surface treatment method of a figurine according to claim 3, characterized in that: After completing multi-layer spraying, the final surface image of the action figure model is obtained based on optical scanning technology, and the surface texture distribution characteristics are extracted through image segmentation algorithm.

5. The surface treatment method of a figurine model according to claim 4, characterized in that: It is determined whether a second preset quality condition is met based on the texture distribution characteristics. If not, a local repair process is used to perform secondary processing on the specific area until the quality requirement is met.

6. The surface treatment method of a figurine model according to claim 1, characterized in that: The geometric feature parameters extracted by the intelligent algorithm include curvature, normal vector direction and boundary contour information.

7. The surface treatment method of a figurine model according to claim 1, characterized in that: The calculation formula of the matching degree is as follows: Among them, M represents the matching degree, S i represents the similarity score of the i-th texture sample, w i Represents the weight coefficient of the i-th texture sample.

8. The surface treatment method of a figurine according to claim 3, characterized in that: The spraying thickness deviation Δh and the spraying uniformity deviation Δu are monitored by an ultrasonic sensor and an optical sensor respectively.

9. The surface treatment method of a figurine according to claim 4, characterized in that: The image segmentation algorithm adopts a method based on edge detection and region growing to distinguish different texture regions and extract their feature parameters.

10. The surface treatment method of a figurine model according to claim 5, characterized in that: The local repair process includes locally spraying the surface of the hand model with a high-precision airbrush, finely polishing the sprayed area with a micro-polishing tool, and optically scanning and quality assessment again.

Images