Correlating Parameters Evaluating Sludge Dewaterability and Morphological Characteristics of Sludge Flocs by a Commercial Smartphone and Image Analysis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The introduction is not well organized, and the connections between its sections need to be strengthened. Furthermore, the contributions of this work are not sufficiently highlighted.

figure 2, in each picture, use the PAM concentration instead of -1, -2 ...

there's an extrac character in line 195, please check.

Is there any possible solution to process black colored water samples?

Author Response

Reviewers’ comments:

Reviewer #1 :The introduction is not well organized, and the connections between its sections need to be strengthened. Furthermore, the contributions to this work are not sufficiently highlighted.

Reply: Thanks for the reviewer’s valuable comments. The connections between sections in the introduction have been strengthened and the previous contributions to this work have been highlighted by adding more sentences in the introduction section as follows:

Floc morphology provides additional information regarding the properties of sludge and serves as an indicator for sludge dewatering or settling (Nakaya et al., 2024). (Page 2, Lines 59-60)

Conventional microscopic analysis presented practical limitations for routine monitoring due to excessive time requirements and substantial operational costs (Li et al., 2019), and a novel technology with less time requirement and easier operation needs to be developed. (Page 3, Lines 76-79)

Currently, limited studies have investigated the diagnostic potential of smartphone cameras for assessing sludge floc morphology and its correlation with dewatering performance. (Page 3, Lines 89-91)

 

1. Figure 2, in each picture, use the PAM concentration instead of -1, -2 ...

Reply: Thanks very much for the reviewer’s valuable comments. We have made the below revision:

 

Figure 2. The original sludge photo for different PAM concentrations (0, 0.05, 0.1, 0.2, 0.3, and 0.5 wt%) in different samples: (a) WAS1, (b) WAS2, and (c) WAS3; processed photographic results of sludge treated at different PAM concentrations in different samples: (d) WAS1, (e) WAS2, and (f) WAS3. (Page 7)

 

2. There's an extra character in line 195, please check.

Reply: Thanks for the reviewer’s valuable comments. The sentence has been revised as follows:

This reduced repulsion between particles allowed them to aggregate (Wang et al., 2025b). (Page 5, Lines 202-203)

 

3. Is there any possible solution to process black colored water samples?

Reply: Thanks for the reviewer’s valuable comments. To maintain the unified parameters of image processing for three different sludge samples, this study failed to process black colored sludge samples. In the future work, sludge samples with more black colors can be processed by modifying the contrast via setting specific code in Matlab software. More sentences have been added as follows:

To ensure consistency in image processing parameters across three types of sludge samples, a unified threshold value was applied during binarization. However, this approach resulted in the failure to effectively process black sludge samples due to the lower contrast between particles and the background hindered accurate segmentation. Although preliminary adjustments indicated that increasing the threshold value could improve segmentation for these black sludge samples, threshold modification was not implemented in this work. This decision was made to avoid introducing variability in image processing, which could affect the comparability of results among the three sludge types. As a result, the black-colored sludge was processed under the same threshold conditions as the other samples, even though this led to less accurate extraction of particle features in those specific cases. In future work, customized threshold adjustment in Matlab could be employed to improve the segmentation performance for black sludge samples, thereby expanding the applicability and accuracy of the image analysis framework. (Page 9, Lines 258-272)

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This study focused on an image-based approach named Fractal image analysis to evaluate sludge dewaterability performance and tried to build a correlation between sludge floc area and key dewaterability parameters. The study is very concise and needs to show the novelty of the study, and write a discussion on the results technically.

1. Abstract: The Abstract is relevant to the study.
2. The literature review does not explicitly articulate what gap is covered in this study. The authors should show the novelty of the current study.
3. Methodology: Add statistical analysis, like ANOVA, and a regression model for this image-based analysis.
4. Results and Discussion: Elaborate discussion on implications of findings and delve into findings within broader applications. Discuss the results, including numerical values. For example, section 3.1 presents a very general discussion. No technical perspective is there.
5. Conclusion: Point 3 should not be a part of the conclusion. Instead, try to focus on the outcome of the study.
6. Address grammar/style issues. For example, line 171-172, “As the dose of cationic PAM increased, the median particle size all showed an increasing trend, regardless of WAS1, WAS2, and WAS3.” is not grammatically correct.

Author Response

Reviewer #2: This study focused on an image-based approach named Fractal image analysis to evaluate sludge dewaterability performance and tried to build a correlation between sludge floc area and key dewaterability parameters. The study is very concise and needs to show the novelty of the study and write a discussion on the results technically.

Reply: Thanks for the reviewer’s positive comments on our study. In the revised version, the novelty of this study has been highlighted, and more discussion of the results has been added. Details can be found in the below sections.  

1. Abstract: The Abstract is relevant to the study.

Reply: Thanks for the reviewer’s positive comments on our abstract section.

 

2. The literature review does not explicitly articulate what gap is covered in this study. The authors should show the novelty of the current study.

Reply: Thanks for the reviewer’s valuable comments. More sentences have been added to show the novelty of this study as follows:
Conventional microscopic analysis presented practical limitations for routine monitoring due to excessive time requirements and substantial operational costs (Li et al., 2019), and a novel technology with less time requirement and easier operation needs to be developed. (Page 3, Lines 76-79)

Currently, limited studies have investigated the diagnostic potential of smartphone cameras for assessing sludge floc morphology and its correlation with dewatering performance. (Page 3, Lines 89-91)

 

3. Methodology: Add statistical analysis, like ANOVA, and a regression model for this image-based analysis.

Reply: Thanks for the reviewer’s valuable comments. More description of the statistical analysis has been added as follows:

2.5. Statistical analysis

Statistical analyses were performed using various software packages. One-way analysis of variance (ANOVA) was conducted using SPSS Statistics 27 for Windows (IBM Corporation, New York, USA), with the least significant difference (LSD) test applied for group comparisons. Prior to ANOVA, data normality was assessed using the Duncan method. Statistical significance was set at p < 0.05, and letter notation was used to indicate differences. Pearson’s correlation coefficients (r) were calculated using SPSS to assess the relationships between sludge floc area and sludge physicochemical properties reflecting sludge dewaterability. Spearman rank correlation analysis was conducted via R software to evaluate the association between sludge floc area and dewatering performance. (Page 5, Lines 162-171)

 

4. Results and Discussion: Elaborate discussion on implications of findings and delve into findings within broader applications. Discuss the results, including numerical values. For example, section 3.1 presents a very general discussion. No technical perspective is there.

Reply: Thanks for the reviewer’s valuable comments. More discussion has been added as follows:

Jin et al. (2016) reported free water increased up to 95.04% by the addition of cationic PAM. While Kopp and Dichtl (2000) reported that polymer conditioning increased the velocity of sludge water release, but the free water content was not influenced by this process. As to the bound water, Katsiris and Kouzeli-Katsiri (1987) reported that decreased bound water content with increased polymer dose. However, Smollen (1990) had a reverse conclusion that the bound water content increased after polymer addition due to the adsorption of water molecules onto the polymer molecules. Wu et al. (1997) suggested that the bound water of alum sludge decreased firstly and then increased with the polymer dose. Chu and Lee (1999) agreed with the findings of Wu et al. (1997), although some argued that bound water was not changed by conditioning (Tsang and Vesilind, 1990). PAM primarily acted by bridging particles to form larger flocs, but this did not always correlate directly with bound water release (Wei et al., 2025). (Page 5, Lines 183-195)

 

5. Conclusion: Point 3 should not be a part of the conclusion. Instead, try to focus on the outcome of the study.

Reply: Thanks for the reviewer’s comments. Point 3 has been deleted. The conclusion has been re-written as follows:

This study highlighted the feasibility of using commercially available smartphones as a practical tool for evaluating sludge dewaterability. The following conclusions can be put forward:

1. Significant correlations were observed between sludge floc area and key dewaterability parameters: The number of flocs in area rangeof10⁻⁶–10⁻⁵ cm² showed a negative correlation with capillary suction time (CST) (regression coefficient (R) = −0.511, probability (p) < 0.05) and a positive correlation with median particle size (R = 0.470, p < 0.05); the number of flocs in area range of 10⁻⁵–10⁻⁴ cm² exhibited a stronger negative correlation with CST (R = −0.538, p < 0.05) and a positive correlation with median particle size (R = 0.480, p < 0.05);
2. When the proportion of the number of flocsin arearange of 10⁻⁵–10⁻⁴ cm² relative to the total floc of conditioned sludge fell below 70%, the dewatered sludge cake achieved a water content of less than 60%. (Page 11, Lines 323-335)

 

6. Address grammar/style issues. For example, line 171-172, “As the dose of cationic PAM increased, the median particle size all showed an increasing trend, regardless of WAS1, WAS2, and WAS3.” is not grammatically correct.

Reply: Thanks for the reviewer’s valuable comments. All grammar mistakes have been checked throughout the manuscript. The sentences in lines 171-172 have been revised as follows:

The results of particle size further corroborate the results of CST and SRF. As the dose of cationic PAM increased, the median particle sizes all showed increased trends, regardless of WAS1, WAS2, and WAS3 samples applied (Figure 1d). (Page 5, Lines 196-198)

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript presents an innovative study that explores the feasibility of using smartphone-based imaging and image analysis techniques to assess sludge dewaterability in wastewater treatment processes. The authors successfully demonstrate correlations between floc morphological characteristics—specifically floc area—and key dewatering performance indicators. Overall, the manuscript is well-structured, with results that are relevant to both academic research and practical applications in WWTPs.

I recommend minor revision before publication. Below are detailed comments to help improve the manuscript’s clarity and completeness.

1. Which statistical test was used to compare the average values between different PAM concentrations in Figure 1? This information should be included in the manuscript.
2. Line 209: The statement “when the floc was circular, the slope shall equal to 1” may be inaccurate. Shouldn’t the slope be equal to 2 in this case? According to the relationship A∝P2/DA \propto P^{2/D}A∝P2/D, for a perfect circle the expected slope in a log–log plot of area versus perimeter would be 2 (since D=1D = 1D=1). Additionally, the cited article (Lee, 2004) does not appear to explicitly support this statement. Please verify and clarify this point.
3. Standard deviations should be added to the results presented in Table 2 to indicate data variability.
4. The manuscript acknowledges challenges associated with darker sludge samples (e.g., WAS2), which reduced the accuracy of image analysis. A more in-depth discussion on how background coloration, lighting, and contrast adjustments influence the analysis. Potential improvements could also be briefly suggested.
5. While the authors correctly report that bound water content did not show consistent trends across sludge types, this point deserves a deeper mechanistic explanation. Why is bound water less reliable in PAM-conditioned sludge?
6. The validation experiments using a laboratory-scale filter press provide a valuable link between image-derived metrics and actual dewatering outcomes. However, since WAS2 was excluded due to image quality limitations, the generalizability of the <70% floc area threshold should be discussed more explicitly. Is this value likely to hold for other sludge types and treatment conditions?
7. Ensure a high-resolution version of Figure 2 is used in the final manuscript. The figure appears pixelated in the current version.
8. The references are not cited numerically in the text, but they are organized numerically in the references section, is that correct? I believe this makes it more difficult for the reader to locate the corresponding references.

Author Response

Reviewer #3: This manuscript presents an innovative study that explores the feasibility of using smartphone-based imaging and image analysis techniques to assess sludge dewaterability in wastewater treatment processes. The authors successfully demonstrate correlations between floc morphological characteristics — specifically floc area—and key dewatering performance indicators. Overall, the manuscript is well-structured, with results that are relevant to both academic research and practical applications in WWTPs. I recommend minor revision before publication. Below are detailed comments to help improve the manuscript’s clarity and completeness.

Reply: Thanks for the reviewer’s positive comments. We have addressed the comments below.

1. Which statistical test was used to compare the average values between different PAM concentrations in Figure 1? This information should be included in the manuscript.

Reply: Thanks for the reviewer’s valuable comments. The method description of statistical test in Figure 1 has been added as follows:

Statistical analyses were performed using various software packages. One-way analysis of variance (ANOVA) was conducted using SPSS Statistics 27 for Windows (IBM Corporation, New York, USA), with the least significant difference (LSD) test applied for group comparisons. Prior to ANOVA, data normality was assessed using the Duncan method. Statistical significance was set at p < 0.05, and letter notation was used to indicate differences. (Page 5, Lines 163-168)

2. Line 209: The statement “when the floc was circular, the slope shall equal to 1” may be inaccurate. Shouldn’t the slope be equal to 2 in this case? According to the relationship A∝P2/DA \propto P^{2/D}A∝P2/D, for a perfect circle the expected slope in a log–log plot of area versus perimeter would be 2 (since D=1D = 1D=1). Additionally, the cited article (Lee, 2004) does not appear to explicitly support this statement. Please verify and clarify this point.

Reply: Thanks for the reviewer’s valuable comments. Our previous statement was correct. If converting this formula into a log form, the equation of “logP_s = D/2 ​log​A_s + constant number” can be obtained. Then the slope shall be D/2, and from the literature (Lee, 2004), the fractal dimension for a circle shall be 2 and the slope shall be 1. There is a difference between D and slope concepts.

 

3. Standard deviations should be added to the results presented in Table 2 to indicate data variability.

Reply: Thanks for the reviewer’s valuable comments. The standard deviations have been added in Table 2 as follows:

Table 2. The number of sludge flocs with different area ranges.

Samples	PAM concentration %	10–6–10–5cm2		10–5–10–4cm2		10–4–10–3cm2		> 10–3 cm2		> 10–6 cm2
		Number	Percentage (%)	Number	Percentage (%)	Number	Percentage (%)	Number	Percentage (%)	Number
WAS1	0.00	75 ± 2	28.3 ± 0.8	181 ± 1	68.3 ± 0.4	8 ± 0	3.0 ± 0.0	0 ± 0	0.0 ± 0.0	265 ± 2.2
	0.05	95 ± 1	29.1 ± 0.3	198 ± 3	60.7 ± 0.9	30 ± 1	9.2 ± 0.3	2 ± 0	0.6 ± 0.0	326 ± 0.5
	0.10	104 ± 5	29.3 ± 1.4	228 ± 3	64.2 ± 0.8	22 ± 2	6.2 ± 0.6	0 ± 0	0.0 ± 0.0	355 ± 0.4
	0.20	137 ± 3	29.4 ± 0.6	285 ± 2	61.2 ± 0.4	43 ± 1	9.2 ± 0.2	0 ± 0	0.0 ± 0.0	466 ± 3.6
	0.30	174 ± 6	28.6 ± 1.0	354 ± 4	58.2 ± 0.7	79 ± 0	13.0 ± 0.0	0 ± 0	0.0 ± 0.0	608 ± 1.7
	0.50	768 ± 10	22.4 ± 0.3	1927 ± 15	56.3 ± 0.4	698 ± 6	20.4 ± 0.2	31 ± 1	0.9 ± 0.0	3425 ± 6.2
WAS2	0.00	3 ± 0	17.7 ± 0.0	12 ± 1	70.6 ± 5.9	2 ± 0	11.8 ± 0.0	0 ± 0	0.0 ± 0.0	17 ± 1.5
	0.05	3 ± 0	8.1 ± 0.0	30 ± 1	81.1 ± 2.7	4 ± 0	10.8 ± 0.0	0 ± 0	0.0 ± 0.0	37 ± 0.5
	0.10	15 ± 1	23.4 ± 1.5	46 ± 0	71.2 ± 0.0	3 ± 0	4.7 ± 0.0	0 ± 0	0.0 ± 0.0	64 ± 0.3
	0.20	3 ± 0	14.3 ± 0.0	16 ± 1	76.2 ± 4.8	2 ± 0	9.5 ± 0.0	0 ± 0	0.0 ± 0.0	21 ± 0.4
	0.30	15 ± 1	44.1 ± 2.9	18 ± 1	52.9 ± 2.9	1 ± 0	2.9 ± 0.0	0 ± 0	0.0 ± 0.0	34 ± 2.3
	0.50	26 ± 1	32.5 ± 1.2	44 ± 2	55.0 ± 2.5	10 ± 1	12.5 ± 1.3	0 ± 0	0.0 ± 0.0	80 ± 1.2
WAS3	0.00	14 ± 1	15.6 ± 1.1	72 ± 1	80.0 ± 1.1	4 ± 0	4.4 ± 0.0	0 ± 0	0.0 ± 0.0	90 ± 0.9
	0.05	23 ± 1	17.6 ± 0.8	89 ± 1	67.9 ± 0.8	19 ± 1	14.5 ± 0.8	0 ± 0	0.0 ± 0.0	131 ± 1.2
	0.10	271 ± 8	31.1 ± 0.9	525 ± 6	60.3 ± 0.7	73 ± 3	8.4 ± 0.3	2 ± 0	0.2 ± 0.0	871 ± 2.3
	0.20	1063 ± 10	41.4 ± 0.4	1384 ± 11	54.0 ± 0.4	115 ± 6	4.5 ± 0.2	3 ± 0	0.1 ± 0.0	2565 ± 1.2
	0.30	458 ± 3	44.0 ± 0.3	548 ± 7	52.7 ± 0.7	33 ± 0	3.2 ± 0.0	1 ± 0	0.1 ± 0.0	1040 ± 2.6
	0.50	138 ± 2	32.1 ± 0.5	246 ± 3	57.2 ± 0.7	46 ± 1	10.7 ± 0.2	0 ± 0	0.0 ± 0.0	430 ± 4.5

(Page 8)

 

4. The manuscript acknowledges challenges associated with darker sludge samples (e.g., WAS2), which reduced the accuracy of image analysis. A more in-depth discussion on how background coloration, lighting, and contrast adjustments influence the analysis. Potential improvements could also be briefly suggested.

Reply: Thank you very much for your suggestions. A more in-depth discussion and potential improvements have been added as follows:

To ensure consistency in image processing parameters across three types of sludge samples, a unified threshold value was applied during binarization. However, this approach resulted in the failure to effectively process black sludge samples due to the lower contrast between particles and the background hindered accurate segmentation. Although preliminary adjustments indicated that increasing the threshold value could improve segmentation for these black sludge samples, threshold modification was not implemented in this work. This decision was made to avoid introducing variability in image processing, which could affect the comparability of results among the three sludge types. As a result, the black colored sludge was processed under the same threshold conditions as the other samples, even though this led to less accurate extraction of particle features in those specific cases. In future work, customized threshold adjustment in Matlab could be employed to improve the segmentation performance for black sludge samples, thereby expanding the applicability and accuracy of the image analysis framework. (Page 9, Lines 258-272)

 

5. While the authors correctly report that bound water content did not show consistent trends across sludge types, this point deserves a deeper mechanistic explanation. Why is bound water less reliable in PAM-conditioned sludge?

Reply: Thanks for the reviewer’s valuable comments. More sentences have been added as follows to discuss the different trends of bound water:

Jin et al. (2016) reported free water increased up to 95.04% by the addition of cationic PAM. While Kopp and Dichtl (2000) reported that polymer conditioning increased the velocity of sludge water release, but the free water content was not influenced by this process. As to the bound water, Katsiris and Kouzeli-Katsiri (1987) reported that decreased bound water content with increased polymer dose. However, Smollen (1990) had a reverse conclusion that the bound water content increased after polymer addition due to the adsorption of water molecules onto the polymer molecules. Wu et al. (1997) suggested that the bound water of alum sludge decreased firstly and then increased with the polymer dose. Chu and Lee (1999) agreed the findings of Wu et al. (1997), although some argued that bound water was not changed by conditioning (Tsang and Vesilind, 1990). PAM primarily acted by bridging particles to form larger flocs, but this did not always correlate directly with bound water release (Wei et al., 2025). (Page 5, Lines 183-195)

 

6. The validation experiments using a laboratory-scale filter press provide a valuable link between image-derived metrics and actual dewatering outcomes. However, since WAS2 was excluded due to image quality limitations, the generalizability of the <70% floc area threshold should be discussed more explicitly. Is this value likely to hold for other sludge types and treatment conditions?

Reply: Thanks for the reviewer’s valuable comments. More discussion about the threshold value of less than 70% and its relationship with practical application for different types of sludge has been added as follows:

Meanwhile, it should be noted that since WAS2 was excluded for data process due to image quality limitations, the generalizability of the <70% floc area for other types of sludge and treatment conditions and its link with sludge dewaterability in real application should be further studied in the future work. (Page 10, Lines 314-318)

 

7. Ensure a high-resolution version of Figure 2 is used in the final manuscript. The figure appears pixelated in the current version.

Reply: Thanks for the reviewer’s valuable comments. The figure has been modified as follows:

 

Figure 4. Relationship between sludge dewatering parameters and the proportion of the floc numbers in each sludge floc area (10^–6-10^–5, 10^–5-10^–4, 10^–4-10^–3 cm², > 10^–3 cm², > 10^–6 cm²) to the total number of flocs. * denotes p < 0.05, and ** denotes p < 0.01.  (Page 9)

8. The references are not cited numerically in the text, but they are organized numerically in the references section, is that correct? I believe this makes it more difficult for the reader to locate the corresponding references.

Reply: Thanks for the reviewer’s valuable comments. The format of references has been unified.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thanks for addressing the comments.